Home |
Search |
Today's Posts |
#81
|
|||
|
|||
Doppler Distoriton?
|
#82
|
|||
|
|||
Doppler Distoriton?
"Goofball_star_dot_etal" wrote in message
On Tue, 10 Aug 2004 19:03:24 GMT, ow (Goofball_star_dot_etal) wrote: Well s**t, (plain english, note) the agreement between my simple-simon model and experiment is within 20% error.. I should worry about blowing smoke! ^^^^^ ^^^^^ Now how do we measure displacement more accurately? The modern way involves a laser and $$$$$$$$ |
#84
|
|||
|
|||
Doppler Distoriton?
"Goofball_star_dot_etal" wrote in message
On Tue, 10 Aug 2004 19:03:24 GMT, ow (Goofball_star_dot_etal) wrote: Well s**t, (plain english, note) the agreement between my simple-simon model and experiment is within 20% error.. I should worry about blowing smoke! ^^^^^ ^^^^^ Now how do we measure displacement more accurately? I did some fiddling with a dial gauge and came up with about 110 thousandths. |
#85
|
|||
|
|||
Doppler Distoriton?
|
#86
|
|||
|
|||
Doppler Distoriton?
"Paul Guy" wrote in message
On Thu, 5 Aug 2004 22:19:23 -0400, "Arny Krueger" wrote: Here's the results of some speaker measurements that I made tonight, based on passing 50 Hz & 4 KHz mixed 1:1 at about 1.2 volts rms, through a Peerless 6.5 inch woofer with about 6 mm Xmax (relatively large for a woofer its size). The speaker is mounted in a roughly 0.4 cubic foot box with no vent. The power amp is a QSC USA 850. This is not very loud. The mic is an ECM8000 that is a few inches from the woofer cone. http://www.pcavtech.com/techtalk/doppler/ The first graph shows the broadband response. The large spikes at 50 Hz and 4 KHz are clearly visible. The second and third harmonics of the 50 Hz tone are about 30 dB down. The spike for the 4 KHz tone is about 5 dB higher than the spike for 50 Hz because the woofer is simply that much more efficient at 4 KHz. The second graph is taken from the same test, with the frequency scale enlarged to show about 400 Hz on either side of 4 KHz. The first pair of large spikes are about 50 Hz on either side of 4 KHz, the second are about 100 Hz on either side of 4 KHz, and so on. The distortion products are probably a mixture of AM and FM distortion, with FM predominating, as the test is contrived to focus on FM. While I've got this set up, any other data that anyone would find interesting? This is a bit late in the thread but..... If you are trying to discriminate FM from AM, you can look at the envelope. AM (IM) distortion should have a typical modulation pattern. In fact we get a mixture of the two. You might need to use some electronics to AM detect the envelope. ie., bandpass filter for 4000 Hz, envelope detector, then watch the AC component at the output, just like a typical AM receiver. I've done this on a PC using the filters in Audition/CEP For the FM component, clip the signal, put the resultant in a 4 khz bandpass, or just measure the sidebands around 4 KHz. For small modulation indices you get only two sidebands for FM (that's the formal definition of NBFM) just like AM. And, if the modulating signal is distorted, you get multiple sidebands with AN, Just like FM. The only difference between AM (IM) and FM (phase) modulation is the phase of the two sidebands. AM they are both in phase (they add to the overall amplitude). In FM they are out of phase, thus they have no effect on the amplitude. The individual amplitudes are the same between AM and FM, assuming the FM is of small deviation (yours is not all that small). Right, but measureing phase in acoustical tests can be challenging. The poster with the nickname of Goofball has some relevant tools at his disposal. I've also discovered that that you can get some clues about AM versus FM with a triple tone test. If the modulating signal is the same for two tones in a doppler test or in the case of jitter, the tone at the higher frequency has the higher modulation index and therefore a different sideband configuration. wc=carrier freq (rad/sec) wm=modulating freq (rad/sec) A=amplitude M=FM modulation index u=AM modulation factor FM: A*cos(wc*t) + 1/2*M*A[ cos(wc + wm)*t - cos(wc - wm)*t] AM: A*cos(wc*t) + 1/2*u*A[ cos(wc + wm)*t + cos(wc - wm)*t] This applies to narrow band FM (M less than 0.3). Notice that the sign of the lower sideband is the only difference. Agreed. A standard spectrum analyzer will not allow you to tell the AM from FM sidebands (to the best of my limited knowledge). Not directly. There might be some tricks you can play if you have a tracking generator. If your spectrum analyzer allows you to sample the data and run math on it, you might be able to massage the data to give you some useful results. That's roughly the approach that is being taken. If your existing spectrum has unequal sidebands, you can suspect a combination of AM and FM sidebands. Agreed. We've got 'em One trick you might pull off, is to ADD pure AM modulated signal to your measured signal. If there is FM modulation, you can add enough AM such that one of the sidebands cancels out. By knowing the AM amplitude, this will allow you to determine the FM component. We've already got a mixture of AM & FM coming out of the test. |
#87
|
|||
|
|||
Doppler Distoriton?
ow (Goofball_star_dot_etal) writes:
On Tue, 10 Aug 2004 19:03:24 GMT, ow (Goofball_star_dot_etal) wrote: Well s**t, (plain english, note) the agreement between my simple-simon model and experiment is within 20% error.. I should worry about blowing smoke! ^^^^^ ^^^^^ Now how do we measure displacement more accurately? An accelerometer, a la Velodyne? -- % Randy Yates % "She's sweet on Wagner-I think she'd die for Beethoven. %% Fuquay-Varina, NC % She love the way Puccini lays down a tune, and %%% 919-577-9882 % Verdi's always creepin' from her room." %%%% % "Rockaria", *A New World Record*, ELO http://home.earthlink.net/~yatescr |
#88
|
|||
|
|||
Doppler Distoriton?
On Wed, 11 Aug 2004 02:57:26 GMT, Randy Yates wrote:
An accelerometer, a la Velodyne? -- And those who are not asleep will note that this is a very good solution from an experimental point of view. Points to those that say why. |
#89
|
|||
|
|||
Doppler Distoriton?
"Goofball_star_dot_etal" wrote in message
On Wed, 11 Aug 2004 02:57:26 GMT, Randy Yates wrote: An accelerometer, a la Velodyne? And those who are not asleep will note that this is a very good solution from an experimental point of view. Points to those that say why. It gives you acelleration, forthwith. Integrate it (essentially pass it through a low pass filter) once and you get velocity. Integrate it again and you get position. Every time you integrate it, a lot of spurious noise gets low-pass filtered @ 6 dB/octave. |
#90
|
|||
|
|||
Doppler Distoriton?
On Wed, 11 Aug 2004 15:28:45 -0400, "Arny Krueger"
wrote: "Goofball_star_dot_etal" wrote in message On Wed, 11 Aug 2004 02:57:26 GMT, Randy Yates wrote: An accelerometer, a la Velodyne? And those who are not asleep will note that this is a very good solution from an experimental point of view. Points to those that say why. It gives you acelleration, forthwith. Integrate it (essentially pass it through a low pass filter) once and you get velocity. Integrate it again and you get position. Every time you integrate it, a lot of spurious noise gets low-pass filtered @ 6 dB/octave. OK, 2 points but remember what Svante said. . and there is more. . . Think "one variable at a time" and "how should we record and process the accelerometer data?" |
#91
|
|||
|
|||
Doppler Distoriton?
ow (Goofball_star_dot_etal) writes:
On Wed, 11 Aug 2004 15:28:45 -0400, "Arny Krueger" wrote: "Goofball_star_dot_etal" wrote in message On Wed, 11 Aug 2004 02:57:26 GMT, Randy Yates wrote: An accelerometer, a la Velodyne? And those who are not asleep will note that this is a very good solution from an experimental point of view. Points to those that say why. It gives you acelleration, forthwith. Integrate it (essentially pass it through a low pass filter) once and you get velocity. Integrate it again and you get position. Every time you integrate it, a lot of spurious noise gets low-pass filtered @ 6 dB/octave. Yes, and that's a good thing(TM). OK, 2 points but remember what Svante said. . and there is more. . . Think "one variable at a time" and "how should we record and process the accelerometer data?" What's the bandwidth of an accelerometer signal when stuck on a speaker cone? Could you use a sound card input to digitize it? -- Randy Yates Sony Ericsson Mobile Communications Research Triangle Park, NC, USA , 919-472-1124 |
#92
|
|||
|
|||
Doppler Distoriton?
On 10 Aug 2004 15:26:56 -0400, Randy Yates
wrote: (Goofball_star_dot_etal) writes: On 10 Aug 2004 10:19:38 -0400, Randy Yates wrote: (Goofball_star_dot_etal) writes: On Tue, 10 Aug 2004 02:39:51 GMT, Randy Yates wrote: Wow, that's eloquent! Hey, any way you can get the idea across! Actually the idea of superposition goes all the way out to abstract mathematics. You can consider a "system" a mapping, call it f(.), from one domain (the input), call it X, to another domain (the output), call it Y. Thus Y = f(X). In abstract algebra, a mapping is "homomorphic" if, for two inputs x1 and x2, f(x1 + x2) = f(x1) + f(x2). Note that the operation "+" on the left side of the "=" is in the domain X, while on the right side it is in the domain Y. A "system" (in the engineering sense) obeys the superposition principle if and only if the mapping corresponding to that system is homomorphic. In other words, "homorphism" and "superposition" are the same thing. -- So I take it that you are not a fan of plain english. . . If the concepts involved were just a matter of "plain english" then it probably wouldn't have taken mathematicians, physicists, and engineers centuries to formalize them. I always thought they were slow.. . Also, I have found that, while intuition (which is what I call the level of explanation you were attempting) can be helpful at certain stages of understanding, over-relying on it can lead to errors in analysis. Agreed but lean about intuition: http://www.amazon.co.uk/exec/obidos/...293460-4259040 :It is necessary, if you want to do more than blow smoke in front of mirrors, to formally derive the results one is asserting. -- Feel free to imply that I am an uncouth blowhard. Hey!?! Peace, man! I ain't trying to rain on your parade. I sorta like you. If I'm implying you're an uncouth blowhard, then I'm also implying I am too! There've been way too many times when I ASSUMED my intuition was correct only to CRASH AND BURN! It ain't pretty. I also ain't the sharpest tool in the shed, and if that is what I implied I apologize. I made the low score (C+) in my last class (OK, so it was a graduate engineering course on error correction with some of the sharpest minds from China and India attending), and my first post-high school education was DeVry (definitely more of a "hands-on" school than a theory school). I think the two worlds (theory vs. practice) are both way cool in their own respects, but to actually unite the two is totally way way way cool.... I get a hard-on trying to do that. I've done the analysis, calculations and analysed the experimental results, spotted some circular arguments and other errors. It ain't "rocket science" but where were you during this time with your faultly BS detector? I want some numbers from you now. Sorry sir. I'll try to improve and cough up the goods from now on... -- Randy Yates Sony Ericsson Mobile Communications Research Triangle Park, NC, USA , 919-472-1124 Since I am the result of a liason between Howard Ferstler and a sheep, you should not expect too much of me: All is revealed: http://groups.google.com/groups?safe...&num=100&hl=en Anyhow I have the 'flue and may be dead tomorrow, in which case could you take a look at: http://www.wareing.dircon.co.uk/images/Bingo.jpg and translate it into a formula for Bob Cain. Sorry I've got to die now, byeeee. |
#93
|
|||
|
|||
Doppler Distoriton?
Bob Cain wrote: Well, I've asked for help on the general equation for pressure at a point removed from an ideal piston in an infinite tube as a function of the force applied to the piston that includes the effects of Doppler distortion in alt.sci.physicw and on the moderated group sci.physics.research where the real guns hang out and there has been no answer. What I've found is that any attempt to write the expression from conditions at the interface results in a recursion or infinite regress unless the term included to account for the motion of the piston is set to zero. It's really tricky. So let's look at an argument by reciprocity. Assume an acoustic pulse of any arbitrary shape running down the tube with an ideal pistion (no mass, stiff, infinite compliance) in place. 1. The piston will move exactly in step with the motion of the air molecules as the pulse passes by it. Now let's measure and record the velocity of that piston as the pulse passes by. Next let's mount a voltage to velocity transducer, again ideal with a zero mechanical impedence, on the side of the piston from which the pulse came when we measured it. 2. When we drive that piston so as to reproduce the velocity that was recorded we will get the identical pulse propegating off of it as originally measured. 3. Because air is air, the resulting pressure pulse will be in phase with that velocity and given by p(t) = v(t) * Ra, where Ra is the characteristic impedence or air, and that pressure pulse will be identical to the one that the measured pulse had. Because this should be true with a pulse of any shape it will be true of a supposition of any such pulses which implies that it is true of any signal and is thus a linear transducer with no distortion of any kind. The only thing that needs to be said to complete this is that any arguments relative to velocity can be equally applied to pressure since pressure and velocity in air are in phase and related by a constant of proportionality, Ra, the acoustic impedence of air. The last nail in the coffin of loudspeaker Doppler distortion is to find the flaw in the the intuitive, and highly persuasive, argument that has been used historically and used a lot as a counter to my argument. The flaw just came to me today, finally. Doppler shift is an unquestionable phenomenon, easily proved from first principles. The thing is that it depends on a sound source that is moving with respect to, i.e. _within_ the medium (we assume for this that the listener is stationary with respect to it.) A piston with two frequencies driving it and stationary with respect to the medium otherwise will supposedly show those frequencies modulating each other. The flaw is that the piston is not in any sense moving within the medium, it is moving the medium which is fundamentally different. If you had a little loudspeaker moving back and forth in space with a low frequency while emitting a high frequency you would indeed see the "Doppler distortion" phenomenon but if you move that speaker back and forth with a large plane to which it is attached, the low and high frequencies will add linearly with no distortion. In the latter case it is part of the system moving the air and in the former it is just moving within the air. If you move the entire plane with the sum of the two frequencies, it moves the air the same amount and generates a pressure wave with only those two spectral components and no frequency modulation. Doppler distortion in loudspeakers in indeed a myth justified by faulty intuitive hand waving. Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein |
#94
|
|||
|
|||
Doppler Distoriton?
"Arny Krueger" wrote in message ...
Here's the results of some speaker measurements that I made tonight, based on passing 50 Hz & 4 KHz mixed 1:1 at about 1.2 volts rms, through a Peerless 6.5 inch woofer with about 6 mm Xmax (relatively large for a woofer its size). The speaker is mounted in a roughly 0.4 cubic foot box with no vent. The power amp is a QSC USA 850. This is not very loud. The mic is an ECM8000 that is a few inches from the woofer cone. http://www.pcavtech.com/techtalk/doppler/ The first graph shows the broadband response. The large spikes at 50 Hz and 4 KHz are clearly visible. The second and third harmonics of the 50 Hz tone are about 30 dB down. The spike for the 4 KHz tone is about 5 dB higher than the spike for 50 Hz because the woofer is simply that much more efficient at 4 KHz. The second graph is taken from the same test, with the frequency scale enlarged to show about 400 Hz on either side of 4 KHz. The first pair of large spikes are about 50 Hz on either side of 4 KHz, the second are about 100 Hz on either side of 4 KHz, and so on. The distortion products are probably a mixture of AM and FM distortion, with FM predominating, as the test is contrived to focus on FM. While I've got this set up, any other data that anyone would find interesting? Could you simulate a Leslie speaker by mixing a really huge low frequency sine wave with the regular audio program? |
#95
|
|||
|
|||
Doppler Distoriton?
"Bob Cain" wrote in message
... Doppler shift is an unquestionable phenomenon, easily proved from first principles. The thing is that it depends on a sound source that is moving with respect to, i.e. _within_ the medium (we assume for this that the listener is stationary with respect to it.) A piston with two frequencies driving it and stationary with respect to the medium otherwise will supposedly show those frequencies modulating each other. The flaw is that the piston is not in any sense moving within the medium, it is moving the medium which is fundamentally different. I don't understand. Explain to me how the speaker actually transfers the energy to the air. Maybe you can do that by contrast. Suppose I hit my knuckle on the table. I can't see the table vibrate, but I know it does. The energy moves from the molecules in the table to the molecules in the air. I can live with that. Then I watch the speaker diaphragm move. Sometimes it moves a lot, sometimes a little. At one point during the movement is the sound wave created? I know it pushes air to create a "breeze" but that's not the sound wave. I'm guessing that when the forward movement stops, the pressure goes back to "normal" and the sound wave is formed. Is that how it works or am I still struggling with truly understanding the mechanics of how a speaker works? I know how it works in layman's terms, but I'm fuzzy about the actual mechanics of getting those molecules smacking into each other. |
#96
|
|||
|
|||
Doppler Distoriton?
"Randy Yates" wrote in message
ow (Goofball_star_dot_etal) writes: On Wed, 11 Aug 2004 15:28:45 -0400, "Arny Krueger" wrote: "Goofball_star_dot_etal" wrote in message On Wed, 11 Aug 2004 02:57:26 GMT, Randy Yates wrote: An accelerometer, a la Velodyne? And those who are not asleep will note that this is a very good solution from an experimental point of view. Points to those that say why. It gives you acelleration, forthwith. Integrate it (essentially pass it through a low pass filter) once and you get velocity. Integrate it again and you get position. Every time you integrate it, a lot of spurious noise gets low-pass filtered @ 6 dB/octave. Yes, and that's a good thing(TM). Agreed. OK, 2 points but remember what Svante said. . I can't find that post... and there is more. . . Think "one variable at a time" and "how should we record and process the accelerometer data?" I thought I covered that. What's the bandwidth of an accelerometer signal when stuck on a speaker cone? Depends on the type of acellerometer. The ones commonly used with subwoofers have dominant poles up in the 100's of Hz. The pole is due to the mass of the acellerometer and the compliance of the mounting. Acellerometers used this way do somewhat change the dynamics of the cone. No problemo at 50 Hz. The *pro* methodology involves a laser and $$$$. The Dumax machine does it that way. Could you use a sound card input to digitize it? No doubt, with suitable preamplification. Now, I'm thinking about calibrating the accelerometer on the cone. |
#97
|
|||
|
|||
Doppler Distoriton?
"Detector195" wrote in message
om Could you simulate a Leslie speaker by mixing a really huge low frequency sine wave with the regular audio program? Not in any practical way. Sorry. |
#98
|
|||
|
|||
Doppler Distoriton?
(Detector195) writes:
[...] Could you simulate a Leslie speaker by mixing a really huge low frequency sine wave with the regular audio program? But why??? You can buy a used leslie for just a few hundred bucks and get the Real McCoy(TM). --A 1969 Leslie 147 Owner -- Randy Yates Sony Ericsson Mobile Communications Research Triangle Park, NC, USA , 919-472-1124 |
#99
|
|||
|
|||
Leslies, was Doppler Distoriton?
"Wessel Dirksen" writes:
"Randy Yates" wrote in message ... (Detector195) writes: [...] Could you simulate a Leslie speaker by mixing a really huge low frequency sine wave with the regular audio program? But why??? You can buy a used leslie for just a few hundred bucks and get the Real McCoy(TM). --A 1969 Leslie 147 Owner Aren't real classic Leslies priced through the roof these days? I don't know - I bought mine about 3 years ago off of Ebay for $600 ($200 of which was shipping charges). It had an amp problem which I fixed. I doubt that they're much more expensive now, but I've been inactive in buying/selling the stuff for awhile. Unless you mean actual new leslies - yes those are extremely expensive. ($2000+?) You might try Goff for more info, http://www.goffprof.com/ -- Randy Yates Sony Ericsson Mobile Communications Research Triangle Park, NC, USA , 919-472-1124 |
#100
|
|||
|
|||
Leslies, was Doppler Distoriton?
There is simulated Leslie sound and there is real Leslie sound.
Simulated sounds, just that, simulated. Not bad, but not like a real Leslie. If you search on ebay you can find all sorts of Leslie stuff. Prices vary, and they are HEAVY and expensive to ship. I play the guitar and have always wanted to play through a real Leslie. Never have though. Randy Yates wrote in message ... "Wessel Dirksen" writes: "Randy Yates" wrote in message ... (Detector195) writes: [...] Could you simulate a Leslie speaker by mixing a really huge low frequency sine wave with the regular audio program? But why??? You can buy a used leslie for just a few hundred bucks and get the Real McCoy(TM). --A 1969 Leslie 147 Owner Aren't real classic Leslies priced through the roof these days? I don't know - I bought mine about 3 years ago off of Ebay for $600 ($200 of which was shipping charges). It had an amp problem which I fixed. I doubt that they're much more expensive now, but I've been inactive in buying/selling the stuff for awhile. Unless you mean actual new leslies - yes those are extremely expensive. ($2000+?) You might try Goff for more info, http://www.goffprof.com/ |
#101
|
|||
|
|||
Doppler Distoriton?
Jim Carr wrote: "Bob Cain" wrote in message ... Doppler shift is an unquestionable phenomenon, easily proved from first principles. The thing is that it depends on a sound source that is moving with respect to, i.e. _within_ the medium (we assume for this that the listener is stationary with respect to it.) A piston with two frequencies driving it and stationary with respect to the medium otherwise will supposedly show those frequencies modulating each other. The flaw is that the piston is not in any sense moving within the medium, it is moving the medium which is fundamentally different. I don't understand. Explain to me how the speaker actually transfers the energy to the air. By pushing and pulling on it. The velocity of the piston is transfered to the particles (voxels) of the air in front of it. Then I watch the speaker diaphragm move. Sometimes it moves a lot, sometimes a little. At one point during the movement is the sound wave created? Whenever it moves, as it moves. The air has no choice but to go along with it and that creates a velocity and pressure wave that are in phase and related by the characteristic acoustic impedence of air. When you push on air, it moves and pushes on the air in front of it but with some delay in the transfer. That's what causes the speed of sound. The push propegates outward from this bit of air to the bit in front of it and that's a wave. Same when you pull on it. Does that help? Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein |
#102
|
|||
|
|||
"Bob Cain" wrote in message
... When you push on air, it moves and pushes on the air in front of it but with some delay in the transfer. That's what causes the speed of sound. The push propegates outward from this bit of air to the bit in front of it and that's a wave. Same when you pull on it. Does that help? Sorta. From what you're saying. the *origin* of each individual wave can take place at any point within the throw of the diaphragm. Is that correct? |
#103
|
|||
|
|||
Jim Carr wrote: "Bob Cain" wrote in message ... When you push on air, it moves and pushes on the air in front of it but with some delay in the transfer. That's what causes the speed of sound. The push propegates outward from this bit of air to the bit in front of it and that's a wave. Same when you pull on it. Does that help? Sorta. From what you're saying. the *origin* of each individual wave can take place at any point within the throw of the diaphragm. Is that correct? Not sure even how to define the origin of the wave in those terms. Thanks for that. I just realized that the assumptions which are being made about that are the flaw in the intuitive description of "Doppler distortion." Something that is ocuring dynamically is being described in terms of a static piston in one sense and dynamically in another. That doesn't work. The distance from the piston to the the sensor isn't relevant to the argument if it is riding the wave. In a way, it's effect is being included twice if you do that. That's a no-no that will lead to false prediction. The flaw in the common argument for "Doppler distortion" has proven very elusive but I think that this nails it. It really is subtle which explains why it's been around so long. I added rec.audio.pro to this because it's highly relevant to the thread on this subject that is happening there. Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein |
#104
|
|||
|
|||
"Bob Cain" wrote in message
... Sorta. From what you're saying. the *origin* of each individual wave can take place at any point within the throw of the diaphragm. Is that correct? Not sure even how to define the origin of the wave in those terms. Thanks for that. I just realized that the assumptions which are being made about that are the flaw in the intuitive description of "Doppler distortion." I'm having a hard time envisioning just one wave being started by one thrust of the piston. Maybe if I fully understood that rather than the quite satisfactory "it just does and that's how a speaker works" mentality I'v always had, I could argue intelligently one way or the other. I added rec.audio.pro to this because it's highly relevant to the thread on this subject that is happening there. Great. Feed me to the wolves. Hey, RAP folks: I hold no degrees in electronics, physics, acoustics, etc. I do not work with audio as a profession. I just find the topic interesting and do not purport myself to be an expert. As I noted earlier in the thread, which was not cross-posted, if I seem condescending, it is because I am trying to explain things in simple terms to myself. Since my logic is usually sound, my guess is that a basic premise somewhere is wrong or incomplete, hence the detailed and simplist explanations. With that said, help me out here. I can't get myself away from the assumption that since a speaker diaphragm has a throw of a certain distance, then the waves started by the diaphragm may be started from any point in that throw. As such two waves which are created a certain time apart may end up traveling different distances to reach my stationary ear, thus a Doppler shift. Measurable? I dunno. Discernible to my ear? Probably not. |
#105
|
|||
|
|||
Jim Carr wrote: With that said, help me out here. I can't get myself away from the assumption that since a speaker diaphragm has a throw of a certain distance, then the waves started by the diaphragm may be started from any point in that throw. As such two waves which are created a certain time apart may end up traveling different distances to reach my stationary ear, thus a Doppler shift. Measurable? I dunno. Discernible to my ear? Probably not. To recapitulate, the problem with that intuitive view, which is the whole basis of believing that there is "Doppler distortion" is that it assumes that the distance from the driver is the distance from the instantaneous position of the piston. That's wrong. The distance from the driver, since it is riding the wave it is creating, is the distance from its zero or rest position, the position about which it oscillates. That doesn't change with the nature of the signal unless there is a DC component. If the distance from the driver is not changing, there is no Doppler shift. None of the proposed scenarios which have the face of the driver oscillating about a rest position will produce Doppler shift despite intuition. Whew! Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein |
#106
|
|||
|
|||
"Jim Carr" With that said, help me out here. I can't get myself away from the assumption that since a speaker diaphragm has a throw of a certain distance, then the waves started by the diaphragm may be started from any point in that throw. As such two waves which are created a certain time apart may end up traveling different distances to reach my stationary ear, thus a Doppler shift. ** A time delay or advance is just that - it is not Doppler. Any such delay or advance depends solely on the position of the cone - not its *velocity*. If a cone is displaced by 10mm, that will introduce a time error of 29 uS or a phase shift of 50 degrees at 5 kHz. Any attempt to measure Doppler frequency shifts must allow for this - most have not. .............. Phil |
#107
|
|||
|
|||
"Phil Allison" wrote in message
"Jim Carr" With that said, help me out here. I can't get myself away from the assumption that since a speaker diaphragm has a throw of a certain distance, then the waves started by the diaphragm may be started from any point in that throw. As such two waves which are created a certain time apart may end up traveling different distances to reach my stationary ear, thus a Doppler shift. ** A time delay or advance is just that - it is not Doppler. Any such delay or advance depends solely on the position of the cone - not its *velocity*. If a cone is displaced by 10mm, that will introduce a time error of 29 uS or a phase shift of 50 degrees at 5 kHz. Any attempt to measure Doppler frequency shifts must allow for this - most have not. That's because this time shift, more specifically the time rate of change of this time shift, is the cause of Doppler. |
#108
|
|||
|
|||
Arny,
This information is probably in this thread somewhere, but it has gotten so long and convoluted that it's much easier just to ask: Are you asking whether FM (Doppler) modulation at the high frequency is the ONLY effect that results when that high frequency in addition to a low frequency (purposely left undefined since the actual values depend on a number of factors in the physical setup) are reproduced in the same transducer, or is there some amount of AM modulation as well? "The Ghost" gave me an idea for determining this without requiring any measurement of the instantaneous cone displacement. Perform an FM discrimination of the received (microphone) signal at the high frequency "carrier." Call the discriminated signal m(t). Regenerate a perfect FM signal using the modulating signal m(t) and subtract that from the original signal. The result is the residual modulation on the signal, which could then be AM-detected to determine if AM is present. Three practical issues which must be dealt with come to mind: 1) How to synchronize the regenerated FM carrier amplitude to the original FM amplitude? Easy answer: emit a signal consisting of the high frequency tone alone for a length of time adequate to measure the amplitude. 2) What modulation index, or depth of modulation, should be used in the regenerated FM signal? Said another way, what gain (if any) should be applied to m(t) when regenerating the FM signal? 3) How do you synchronize the regenerated signal in time with the original signal? There are actually two synchronization tasks to be done: phase synchronization of the carriers, and delay in the modulating signal, i.e., tau in A*m(t-tau). (A is the parameter in question 2). Does this make any sense? -- % Randy Yates % "My Shangri-la has gone away, fading like %% Fuquay-Varina, NC % the Beatles on 'Hey Jude'" %%% 919-577-9882 % %%%% % 'Shangri-La', *A New World Record*, ELO http://home.earthlink.net/~yatescr |
#109
|
|||
|
|||
"Randy Yates" wrote in message
Arny, This information is probably in this thread somewhere, but it has gotten so long and convoluted that it's much easier just to ask: Are you asking whether FM (Doppler) modulation at the high frequency is the ONLY effect that results when that high frequency in addition to a low frequency (purposely left undefined since the actual values depend on a number of factors in the physical setup) are reproduced in the same transducer, or is there some amount of AM modulation as well? I'm not asking that question, because I know the answer, and I knew it walking in the door last week. The results of playing multiple tones through something as dirty as a speaker produces copious amonts of both AM and FM. As a rule, the AM dominates. "The Ghost" gave me an idea for determining this without requiring any measurement of the instantaneous cone displacement. Perform an FM discrimination of the received (microphone) signal at the high frequency "carrier." Call the discriminated signal m(t). Regenerate a perfect FM signal using the modulating signal m(t) and subtract that from the original signal. The result is the residual modulation on the signal, which could then be AM-detected to determine if AM is present. I've tried that, and a lot of other things. It has the usual problems with nulling in the real world. You can get roughly a 2:1 to 10:1 reduction of the unwanted distortion by that means. Three practical issues which must be dealt with come to mind: 1) How to synchronize the regenerated FM carrier amplitude to the original FM amplitude? Pretty easy to do an fair job of in the digital domain. Easy answer: emit a signal consisting of the high frequency tone alone for a length of time adequate to measure the amplitude. If you've looked at the raw data page posted at http://www.pcavtech.com/techtalk/doppler/ you'd know that finding that out with pretty fair precision is a matter of reading numbers off a screen. 2) What modulation index, or depth of modulation, should be used in the regenerated FM signal? Said another way, what gain (if any) should be applied to m(t) when regenerating the FM signal? At this point I should point out that since the AM dominates, it might make sense to apply an AM signal to null the AM part out, leaving the FM. 3) How do you synchronize the regenerated signal in time with the original signal? There are actually two synchronization tasks to be done: phase synchronization of the carriers, and delay in the modulating signal, i.e., tau in A*m(t-tau). (A is the parameter in question 2). Well, we know quite a bit about the signal that we are trying to clean up. Does this make any sense? Been there, done that. Seriously, I come back to this problem of separating AM and FM from a real world signal every once and while, and learn a bit more about solving it. This time I realized that ideally, AM distortion related sidebands are indepenendent of of the carrier frequency, but increase in amplitude with carrier frequency for FM. Trouble is, this practical example is so heavily dominated by the AM distortion. I hope to go back to studying jitter, and play this card there. I suspect that lots of people have been misidentifying AM distortion products as jitter. |
#110
|
|||
|
|||
"Phil Allison" wrote in message
... ** A time delay or advance is just that - it is not Doppler. Any such delay or advance depends solely on the position of the cone - not its *velocity*. I disagree. The time delay or advance *is* Doppler. The speed of sound is constant in a given medium. In the classic example of the train whistle the velocity of the train changes the distance/time between waves, but those waves travel just as fast as if the train were still. It seems like you are saying Doppler has to do with adding velocities together, which is untrue. |
#111
|
|||
|
|||
In alt.music.home-studio,rec.audio.tech,rec.audio.pro,
Bob Cain wrote: Jim Carr wrote: With that said, help me out here. I can't get myself away from the assumption that since a speaker diaphragm has a throw of a certain distance, then the waves started by the diaphragm may be started from any point in that throw. As such two waves which are created a certain time apart may end up traveling different distances to reach my stationary ear, thus a Doppler shift. Measurable? I dunno. Discernible to my ear? Probably not. To recapitulate, the problem with that intuitive view, which is the whole basis of believing that there is "Doppler distortion" is that it assumes that the distance from the driver is the distance from the instantaneous position of the piston. That's wrong. The distance from the driver, since it is riding the wave it is creating, is the distance from its zero or rest position, the position about which it oscillates. That doesn't change with the nature of the signal unless there is a DC component. Oh, goody. You want a DC component, I'll GIVE you a DC component! insert emoticon here Let's go to extremes. Say we got one of these big honking high-power woofers (that I recall reading Arny's writings about a while back) with X-max of one or two inches or so. Superimpose a 1kHz tone (probably the highest frequency it will reasonably reproduce) onto a "DC component", say a 1/2 Hz sine wave that slooowly moves the cone in and out a total distance of two inches, all the while it's also putting this 1kHz tone into the air. Don't think of it as a 0.5Hz sine wave, think of it as a varying DC component (that's obviously what it is, you can see the cone move back and forth with your eyes). What will you say is the acoustic source of the 1kHz, the driver frame, which does not move, or the cone, which DOES move? And presuming you see this as a possible cause of doppler distortion, how is this "DC component" any different from a higher frequency (say 20Hz or 50Hz) that also causes substantial cone displacement? If the distance from the driver is not changing, there is no Doppler shift. The distance from which part of the driver? The frame? The cone? Something else? None of the proposed scenarios which have the face of the driver oscillating about a rest position will produce Doppler shift despite intuition. Whew! If you think it's frustrating for you, imagine how I feel with Phil agreeing with me! Bob ----- http://mindspring.com/~benbradley |
#112
|
|||
|
|||
"Jim Carr" "Phil Allison" ** A time delay or advance is just that - it is not Doppler. Any such delay or advance depends solely on the position of the cone - not its *velocity*. I disagree. The time delay or advance *is* Doppler. ** Doppler frequency shift is proportional to source velocity - so they are not the same. The speed of sound is constant in a given medium. ** If the medium is moving at some speed then that adds to, or subtracts from, the speed of sound in still air (ie 343 m/S) In the classic example of the train whistle the velocity of the train changes the distance/time between waves, ** Yes, because the train is moving through the air. but those waves travel just as fast as if the train were still. It seems like you are saying Doppler has to do with adding velocities together, which is untrue. ** Not at all - but the magnitude of the Doppler shift is proportional to the velocity of the source compared to the surrounding air. A woofer cone takes a small volume of with it for the ride. ............. Phil |
#113
|
|||
|
|||
"Phil Allison" wrote in message ... I disagree. The time delay or advance *is* Doppler. ** Doppler frequency shift is proportional to source velocity - so they are not the same. Maybe we just have a failure to communicate. I say the actual frequency does not shift. Assume a constant frequency at the source. If the distance between the source and the receiver changes, then the receiver cannot reliably determine the frequency. This is because the distance between the sound waves (wavelength) emitted by the source changes. The receiver determines the frequency by measuring the wavelength, which is the distance between a given point on a wave and the corresponding point on the next cycle of the wave. Now, our ears don't measure distance, they are measuring time to put it loosely. They don't care how fast the wave is moving. They sense the time interval (a function of distance) between waves. Since movement of either the source, observer or both can change that distance, there is an *apparent* shift in frequency, not a "real" shift. We know this because we already agreed the source emitted a constant frequency. It can be expressed like this: fo = fs . (v - vo) / (v - vs) fo is the apparent frequency of the observer. fs is the frequency of the source v is the speed of sound vo and vs are the velocities of the source and observer. But then again, you seemed to disagree about the speed of sound being constant when you wrote... ** If the medium is moving at some speed then that adds to, or subtracts from, the speed of sound in still air (ie 343 m/S) To which I say, care to cite a reference? The apparent shift is certainly a function of the velocity of the source and/or receiver, but the speed of sound is constant. If it's not, there is no Doppler. Let's just reason it out by using the whistle on the train and the changing speed of sound which you cite. Suppose further that the listener is 3,430 meters away. The sound wave has to travel 10 seconds to arrive at the observer, right? Train #1 is stationary and blows the whistle. A second train is moving at 100m/S. At the exact moment is parallel with train #1 it blows its whistle. This happens to be the exact same time train #1 blew its whistle. Based on your statement above, the first sound of the whistle is moving at 343m/S. The second is traveling at 443m/S. Are you saying that the second whistle will arrive at the receiver 2.25 seconds earlier and that we would hear two separate whistles? If they did, they would sound exactly the same anyway, which would mean no Doppler shift. Refer to the formula above. For Doppler to work, v (speed of sound) must remain constant in the medium. If the velocities add together, then the distance between each wave would therefore remain constant. Hence, no shift. |
#114
|
|||
|
|||
"Arny Krueger" "Phil Allison" "Jim Carr" With that said, help me out here. I can't get myself away from the assumption that since a speaker diaphragm has a throw of a certain distance, then the waves started by the diaphragm may be started from any point in that throw. As such two waves which are created a certain time apart may end up travelling different distances to reach my stationary ear, thus a Doppler shift. ** A time delay or advance is just that - it is not Doppler. Any such delay or advance depends solely on the position of the cone - not its *velocity*. If a cone is displaced by 10mm, that will introduce a time error of 29 uS or a phase shift of 50 degrees at 5 kHz. Any attempt to measure Doppler frequency shifts must allow for this - most have not. That's because this time shift, more specifically the time rate of change of this time shift, is the cause of Doppler. ** So this is what all the Doppler Distortion fuss is about ???? A tiny bit of phase jitter, which at 5 kHz rarely amounts to more than a few degrees ?? I was looking at it on my scope yesterday: 1. A 5 inch woofer, in box, driven by an amp fed from with two sine wave generators with outputs summed. 2. A condenser mic feeding a pre-amp and followed by a 12 dB/oct HPF at 2 kHz thence to the scope. 3. The high frequency generator output is also linked to the scope which operates in X-Y mode. 4. Park mic in front of woofer fed with a circa 5000 Hz sine wave at about 10 watts. ( I used ear muffs) 5. Adjust scope and exact mic position to get a straight, diagonal line traced on the scope screen - note that adjusting the 5000 Hz amplitude affects the angle of the diagonal line only (ie makes it easy to visually distinguish amplitude modulation ). 6. Turn up low frequency generator, set to say 40 Hz, and watch the line open out to form a narrow ellipse indicating that the phase is changing as the cone moves closer and further away from the mic. 7. Sweep low frequency generator up and down and note that cone excursion alone controls the size of the ellipse - it never opens out more than about 15 degrees for a linear cone excursion of 3 mm. 8. Try hard to imagine that this is the notorious, evil, Doppler distortion before your eyes. Wow. ........... Phil |
#115
|
|||
|
|||
"Jim Carr" "Phil Allison" I disagree. The time delay or advance *is* Doppler. ** Doppler frequency shift is proportional to source velocity - so they are not the same. Maybe we just have a failure to communicate. I say the actual frequency does not shift. Since movement of either the source, observer or both can change that distance, there is an *apparent* shift in frequency, not a "real" shift. ** The *observed* frequency changes - end of story. But then again, you seemed to disagree about the speed of sound being constant when you wrote... ** If the medium is moving at some speed then that adds to, or subtracts from, the speed of sound in still air (ie 343 m/S) To which I say, care to cite a reference? ** Silly question - see any school physics text. Let's just reason it out by using the whistle on the train and the changing speed of sound which you cite. Suppose further that the listener is 3,430 meters away. The sound wave has to travel 10 seconds to arrive at the observer, right? Train #1 is stationary and blows the whistle. A second train is moving at 100m/S. At the exact moment is parallel with train #1 it blows its whistle. This happens to be the exact same time train #1 blew its whistle. Based on your statement above, the first sound of the whistle is moving at 343m/S. The second is traveling at 443m/S. ** That wrong conclusion simply has no connection with my statement. Both whistle sounds travel through still air to the listener. Are you saying that the second whistle will arrive at the receiver 2.25 seconds earlier and that we would hear two separate whistles? ** Nope. ............ Phil |
#116
|
|||
|
|||
Arny Krueger wrote: That's because this time shift, more specifically the time rate of change of this time shift, is the cause of Doppler. Doesn't exist, Arny. Look he http://www.silcom.com/~aludwig/Physi..._of_sound.html Tellingly, as deep as the discussion goes, no mention is made of "Doppler distortion" and if you read it you will see why such nonsense wouldn't even have been considered. It also directly supports what I have said recently that distance from an oscilating piston, for the purposes of the physics of piston interaction with air is the distance to the rest position. I must say, that I found this link just minutes ago, oddly enough looking for links to IR's for Acoutic Modeler. I fingered it out earlier all by m'self. Now about that data you posted... Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein |
#117
|
|||
|
|||
Ben Bradley wrote: Let's go to extremes. Say we got one of these big honking high-power woofers (that I recall reading Arny's writings about a while back) with X-max of one or two inches or so. Superimpose a 1kHz tone (probably the highest frequency it will reasonably reproduce) onto a "DC component", say a 1/2 Hz sine wave that slooowly moves the cone in and out a total distance of two inches, all the while it's also putting this 1kHz tone into the air. Don't think of it as a 0.5Hz sine wave, think of it as a varying DC component (that's obviously what it is, you can see the cone move back and forth with your eyes). What will you say is the acoustic source of the 1kHz, the driver frame, which does not move, or the cone, which DOES move? Doesn't matter how slow the oscilation is, it won't produce Doppler shift. If the distance from the driver is not changing, there is no Doppler shift. The distance from which part of the driver? The frame? The cone? Something else? The rest position, the one it will settle to when the driving signal is removed. If the driving signal contains a DC component, and the piston is not restrained by a compliance, then and only then will Doppler shift occurs. Hard to swallow, I know but it is the truth. None of the proposed scenarios which have the face of the driver oscillating about a rest position will produce Doppler shift despite intuition. Whew! If you think it's frustrating for you, imagine how I feel with Phil agreeing with me! That really must be rough. I sympathize. :-) Look here if you aquire the math to understand it. http://www.silcom.com/~aludwig/Physi..._of_sound.html I just found it and everything I've said is in it if not in the same context. It doesn't consider "Doppler distortion" because there is no reason to. Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein |
#118
|
|||
|
|||
"Phil Allison" wrote in message
"Arny Krueger" "Phil Allison" "Jim Carr" With that said, help me out here. I can't get myself away from the assumption that since a speaker diaphragm has a throw of a certain distance, then the waves started by the diaphragm may be started from any point in that throw. As such two waves which are created a certain time apart may end up travelling different distances to reach my stationary ear, thus a Doppler shift. ** A time delay or advance is just that - it is not Doppler. Any such delay or advance depends solely on the position of the cone - not its *velocity*. If a cone is displaced by 10mm, that will introduce a time error of 29 uS or a phase shift of 50 degrees at 5 kHz. Any attempt to measure Doppler frequency shifts must allow for this - most have not. That's because this time shift, more specifically the time rate of change of this time shift, is the cause of Doppler. ** So this is what all the Doppler Distortion fuss is about ???? A tiny bit of phase jitter, which at 5 kHz rarely amounts to more than a few degrees ?? It's not a lot. The most important thing is that its swamped by all teh AM distortion. I was looking at it on my scope yesterday: 1. A 5 inch woofer, in box, driven by an amp fed from with two sine wave generators with outputs summed. 2. A condenser mic feeding a pre-amp and followed by a 12 dB/oct HPF at 2 kHz thence to the scope. 3. The high frequency generator output is also linked to the scope which operates in X-Y mode. 4. Park mic in front of woofer fed with a circa 5000 Hz sine wave at about 10 watts. ( I used ear muffs) 5. Adjust scope and exact mic position to get a straight, diagonal line traced on the scope screen - note that adjusting the 5000 Hz amplitude affects the angle of the diagonal line only (ie makes it easy to visually distinguish amplitude modulation ). 6. Turn up low frequency generator, set to say 40 Hz, and watch the line open out to form a narrow ellipse indicating that the phase is changing as the cone moves closer and further away from the mic. 7. Sweep low frequency generator up and down and note that cone excursion alone controls the size of the ellipse - it never opens out more than about 15 degrees for a linear cone excursion of 3 mm. 8. Try hard to imagine that this is the notorious, evil, Doppler distortion before your eyes. I never said it was notorious or evil. But net it out -we're saying pretty much the same thing, Phil. The Doppler distortion is there but its small. I think the guy who brought up Doppler as some kind of a serious problem did so a few weeks ago. He used Doppler distortion as a justification for not liking long-excursion woofers. In the end he admitted that he used 2-way monitors with either 6.5 or 8" woofers, and no subwoofer. Ironic enough? |
#119
|
|||
|
|||
"Bob Cain" wrote in message
Arny Krueger wrote: That's because this time shift, more specifically the time rate of change of this time shift, is the cause of Doppler. Doesn't exist, Arny. Look he http://www.silcom.com/~aludwig/Physi..._of_sound.html Tellingly, as deep as the discussion goes, no mention is made of "Doppler distortion" and if you read it you will see why such nonsense wouldn't even have been considered. It also directly supports what I have said recently that distance from an oscilating piston, for the purposes of the physics of piston interaction with air is the distance to the rest position. I must say, that I found this link just minutes ago, oddly enough looking for links to IR's for Acoutic Modeler. I fingered it out earlier all by m'self. Now about that data you posted... Sorry Bob, but I'm not buying. |
#120
|
|||
|
|||
"Arny Krueger" "Phil Allison" ** So this is what all the Doppler Distortion fuss is about ???? A tiny bit of phase jitter, which at 5 kHz rarely amounts to more than a few degrees ?? It's not a lot. The most important thing is that its swamped by all the AM distortion. ** Which, unlike the puny phase jitter, is veeerrrryy audible. I was looking at it on my scope yesterday: 1. A 5 inch woofer, in box, driven by an amp fed from with two sine wave generators with outputs summed. 2. A condenser mic feeding a pre-amp and followed by a 12 dB/oct HPF at 2 kHz thence to the scope. 3. The high frequency generator output is also linked to the scope which operates in X-Y mode. 4. Park mic in front of woofer fed with a circa 5000 Hz sine wave at about 10 watts. ( I used ear muffs) 5. Adjust scope and exact mic position to get a straight, diagonal line traced on the scope screen - note that adjusting the 5000 Hz amplitude affects the angle of the diagonal line only (ie makes it easy to visually distinguish amplitude modulation ). 6. Turn up low frequency generator, set to say 40 Hz, and watch the line open out to form a narrow ellipse indicating that the phase is changing as the cone moves closer and further away from the mic. 7. Sweep low frequency generator up and down and note that cone excursion alone controls the size of the ellipse - it never opens out more than about 15 degrees for a linear cone excursion of 3 mm. 8. Try hard to imagine that this is the notorious, evil, Doppler distortion before your eyes. I never said it was notorious or evil. ** I was using poetic license. But net it out -we're saying pretty much the same thing, Phil. The Doppler distortion is there but its small. ** I found with my test that much of it goes away if you put the woofer cone at right angles to the mic. Hey - I have always had my 10 inch sub mounted facing the floor. Way to go !!! I think the guy who brought up Doppler as some kind of a serious problem did so a few weeks ago. He used Doppler distortion as a justification for not liking long-excursion woofers. In the end he admitted that he used 2-way monitors with either 6.5 or 8" woofers, and no subwoofer. Ironic enough? ** Must just love all that IM. .......... Phil |