"TonyP" wrote in message
u...

"Porky" wrote in message
...
It seems that you missed the point, which was that if there is Doppler
shift in a mic because of diaphragm motion, then there must also be
Doppler
shift in the human ear
because of eardrum motion. If that is true then the hearing mechanism
must
have a method of compensation for it.

It seems you've missed my point. The brain "compensates" for the auditory
system itself, because you have NO other point of reference.


Sure you do, ever hear of bone conduction? That bypasses the eardrum
totally. There is also considerable evidence that we can "hear" the audio in
audio modulated RF at certain frequencies, which would seem to bypass the
physical hearing mechanism entirely. However, my point was that if one can
tell the difference between a "live" sound and the same sound reproduced on
a very high quality sound system, then either there is no audible distortion
present, or our ears have a mechanism that compensated for whatever
distortion is present, including any Doppler distortion.

Porky wrote:


If we get into Doppler shift due to motion in air molecules, I suspect
we're getting down to "the bumble bee doesn't really fly because the math
says it can't" point.

Huh? That's exactly what Doppler distortion is due to.
There is a mild non-linear relationship everywhere in a
soundfield between particle velocity about a point and the
fluid velocity at that point. That was my epiphany. Not
much as epiphanies go but, hey, they get fewer and fewer
every year. :-)

If the equations show that all that much Doppler distortion in a speaker,
why can't we hear it?

Who says we can't? It's sorta hard to get rid of to do an
ABX test on.

By Occam's Rasor, either our hearing mechanism has built in compensation,
so Doppler distortion doesn't matter, or the math is wrong and it needs to
be revised. That isn't to say that it doesn't happen in the
piston-in-an-infinite-tube model, it just means that the speaker/room model
is a totally different animal.

Sure it is but it should get signifigantly worse with a
driver in an enclosure rather than in a tube because of the
large excursions demanded at the low frequencies to couple
anything from an enclosed speaker to a room. You can get
real high SPL low frequencies in a tube without much
excursion, which is what causes it, but not so with an
enclosed speaker in a room. You have to push a whole lot of
LF air up close to get what's in the signal to reach you at
any distance.

Dunno what you mean by built in compensation nor why there
would be anything like that. It's not the kind of thing
evolution would have devoted much energy to. There weren't
many broadband sound sources to work with even if it had
been deemed important for some reason.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

Porky wrote:

It seems that you missed the point, which was that if there is Doppler
shift in a mic because of diaphragm motion, then there must also be Doppler
shift in the human ear
because of eardrum motion. If that is true then the hearing mechanism must
have a method of compensation for it.

There is none of _any_ signifigance with either. Excursions
in either case are more than a few orders of magnitude
away from signifigance.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

"Porky" wrote in message
...

"TonyP" wrote in message
It seems you've missed my point. The brain "compensates" for the
auditory
system itself, because you have NO other point of reference.

Sure you do, ever hear of bone conduction? That bypasses the eardrum
totally.

Not at all. It is concurrent with the eardrum. Audioligists must use large
masking signals just to get some figure for bone conduction, but it is less
than via the eardrum. I know of no way to seperate the two for comparison
purposes, do you? (The worlds audiologists await your reply :-)


There is also considerable evidence that we can "hear" the audio in
audio modulated RF at certain frequencies, which would seem to bypass the
physical hearing mechanism entirely.

Some level of diode detection has been demonstated in some cases, usually
connected with metal fillings. This couples audio signals via bone
conduction.

However, my point was that if one can
tell the difference between a "live" sound and the same sound reproduced
on
a very high quality sound system, then either there is no audible
distortion
present, or our ears have a mechanism that compensated for whatever
distortion is present, including any Doppler distortion.

???

TonyP.

[snips]
"zigoteau" wrote in message
om...
Bob Cain wrote in message
...
Mine, which involves no approximation, yields:

Vp(d,t) = Vd(t - (d - Sd(t-d/c))/c)

I am sorry, Bob, but the second equation does involve an approximation.

Zigoteau.

I agree completely, and I would like to point out that if you evaluate the
above equation at d=0 you get an result identical to my equation (8) in my
analysis at
http://www.silcom.com/~aludwig/Physi...on/dopdist.htm, which has
been posted for quite a while now. Rod Elliot presented the same equation
before I did. I present this equation as an approximate analysis, but show
that it is quite accurate. I also show that this equation is completely
equivalent to the good old fashioned "Doppler distortion" that folks have
been using for years. Just follow the link under "interesting issues" for a
proof of this statement. As far as I am aware you are not claiming your
analysis is novel, Zigoteau, so I don't mean this as a personal criticism.
I just want to set the record straight for people not familiar with the
history.
Art Ludwig

"Jim Carr" wrote in message news:JWp8d.12173$mS1.9564@fed1read05...

What bothered me was that I could not (and still cannot) see how a speaker
really works. Yeh, I can describe the mechanics involved, but I still don't
fully understand the exact physics where the diaphragm creates the sound
wave. Is it at the start of the throw? The end? The middle? If it's in the
middle of a long throw for a loud low frequency, how does it make the higher
frequencies at the same time?


That's odd. I don't find this hard to imagine, slowing it down in my
mind. The highest speed of the diaphragm forward will result in the
highest "sweeping up" of air molecules, resulting in the highest local
density. In this mental image, then, the middle of the throw (for a
fundamental tone) would produce the peak of the oscillation in the
medium.

As far as the harmonics, simply remind yourself of the waveform of,
say, a flute note, as seen on an oscilloscope, periodic but not
sinusoidal. According to the above picture, that's the velocity
profile of the diaphragm.

This is perhaps why a square wave is hard to push through a speaker,
because it demands instantaneous accelerations of the diaphragm.

PD

zigoteau wrote:

Bob Cain wrote in message ...

Hi, Bob,

Just a minor correction.

Please notice the difference between what I wrote and what
Zigoteau wrote. His approach, a first order approximation
using a M-T power series, yields (using common symbols and
frames of reference):

Vp(d,t) = Vd(t - (d - Sd(t-d/c))/(c - Vd(t-d/c)))

Mine, which involves no approximation, yields:

Vp(d,t) = Vd(t - (d - Sd(t-d/c))/c)


I am sorry, Bob, but the second equation does involve an approximation.

I was afraid you were going to say something like that. :-)

From the logic I used to get it could you please show me
the trap I've fallen into? It goes as follows:

The fluid velocity (in units of distance/time) of a test
particle (a tiny, zero mass thing) located a distance d from
the driver face is given by

Vp(d,t) = Vd(t-d/c) 1)

Where Vd() is the velocity of the driver. The position of
the particle relative to its rest position is

Sp(d,t) = Sd(t-d/c) 2)

Where Sd() is the position of the driver relative to its
rest position.

What then is the velocity of a second test particle that is
located at the first particle's rest position? The distance
between the two test particles is Sp(d,t) so that the
velocity of the second particle should be given by

Vf(d,t) = Vp(d-Sp(d,t),t) 3)

Where the d on the LHS is the distance from the driver rest
position to the second particle and the d on the right hand
side is the distance from the driver face to the first
particle. These have the same value.

Substituting 1) into 3) gives

Vf(d,t) = Vd(t-(d-Sp(d,t))/c) 4)

Substituting 2) into 4) gives

Vf(d,t) = Vd(t-(d-Sd(t-d/c))/c) 5)

I'm afraid I can't see anything in that chain of logic that
makes 5) an approximate solution and would appreciate some
help with that if you are certain that it is approximate.


Thanks,

Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

"Paul Draper" wrote in message
m...

That's odd. I don't find this hard to imagine, slowing it down in my
mind. The highest speed of the diaphragm forward will result in the
highest "sweeping up" of air molecules, resulting in the highest local
density. In this mental image, then, the middle of the throw (for a
fundamental tone) would produce the peak of the oscillation in the
medium.

Thanks for the feedback. Take me through a simple sine wave. It starts at
the zero line. It does a semicircle above the line an a semicircle below the
line. Describe for me the motion of the speaker in relation to that and at
what point the wave is created.

I will note that even though this is a contentious thread, I am making a
sincere request to help me envision this. Maybe I'm thinking too hard or
have some sort of mental block on this issue.

This is perhaps why a square wave is hard to push through a speaker,
because it demands instantaneous accelerations of the diaphragm.

I always though it was because most speakers are round! Nyuk!

"Jim Carr" wrote in message
news:vbK8d.13387$mS1.11043@fed1read05...
"Paul Draper" wrote in message
m...

That's odd. I don't find this hard to imagine, slowing it down in my
mind. The highest speed of the diaphragm forward will result in the
highest "sweeping up" of air molecules, resulting in the highest local
density. In this mental image, then, the middle of the throw (for a
fundamental tone) would produce the peak of the oscillation in the
medium.

Thanks for the feedback. Take me through a simple sine wave. It starts at
the zero line. It does a semicircle above the line an a semicircle below
the
line. Describe for me the motion of the speaker in relation to that and at
what point the wave is created.

Actually it's an ellipse, not a semicircle, and that does make a difference.
The cone starts from zero at a fairly rapid rate of acceleration, and the
rate slows gradually until it reaches the peak excursion where it reverses
direction and gradually starts accelerating in the other direction, the rate
of acceleration increases to a peak value as it passes through zero again,
then the rate starts decreasing again until it again reverses direction
again at the "bottom" of the cycle, where the rate starts increasing again
until it again reaches zero, then the whole thing repeats again. Due to the
inertia and compressability of air, the air pressure at the cone's surface
will vary in proportion to the cone's speed, and this pressure variation
becomes a sound wave.
Personally, I think the "virtual" source point of the sound wave being
radiated by the speaker in the rest point of the cone at the center of the
excursion limits, but this has been a subject of debate for a long time, and
there are quite a few other theories. One other theory is that the
instantaneous virtual sound source point is the position of the cone at any
given point in time, which is where the notion of Doppler distortion in a
speaker originates. Note that if the virtual source is really the rest point
of the cone, then there is no Doppler shift in a loud speaker because the
virtual source is not moving with respect to the listener. If the latter
theory is correct, then Doppler shift is produced by a loudspeaker because
the virtual source is moving with respect to the listener.


I will note that even though this is a contentious thread, I am making a
sincere request to help me envision this. Maybe I'm thinking too hard or
have some sort of mental block on this issue.

This is perhaps why a square wave is hard to push through a speaker,
because it demands instantaneous accelerations of the diaphragm.

I always though it was because most speakers are round! Nyuk!

Nope, you can push a square wave through a speaker, but you can only do it
once because the sharp edges will tear up the cone. :-)
Seriously, some very good speakers will produce a pretty good approximation
of a square wave at higher frequencies, but none will do it at lower
frequencies, and no speaker will reproduce a true square wave at any
frequency because a reproducing a true square wave would require that the
cone travel instantaneously from one excursion limit to the other, and as
long as the cone has mass, that ain't gonna happen!

"Bob Cain" wrote in message
...


Porky wrote:


If we get into Doppler shift due to motion in air molecules, I suspect
we're getting down to "the bumble bee doesn't really fly because the
math
says it can't" point.

Huh? That's exactly what Doppler distortion is due to.
There is a mild non-linear relationship everywhere in a
soundfield between particle velocity about a point and the
fluid velocity at that point. That was my epiphany. Not
much as epiphanies go but, hey, they get fewer and fewer
every year. :-)

Doppler shift occurs because as the source moves toward the listener, the
source's motion causes the apparent wavelength of the sound to shorten, and
as it moves away it causes the apparent wavelength of the sound to lengthen.
I say "apparent wavelength" because anyone traveling along with the source
hears it as a steady tone, meaning the actual wavelength does not change.
The point I was trying to make (I think) is that soundwaves travel at the
speed of sound but the individual air molecules hardly travel at all.

If the equations show that all that much Doppler distortion in a
speaker,
why can't we hear it?

Who says we can't? It's sorta hard to get rid of to do an
ABX test on.

As I said, some speaker companies have done ABX testing between a
reproduced sound and a live sound source, and under the best conditions,
even expert listeners had trouble relaibly telling the difference. They
weren't trying to test whether Doppler shift existed in a loudspeaker, but
if Doppler shift in a speaker were all that audible, why didn't it clue the
listeners in to which was the speaker and which was live?
If you say that the live source produced Doppler shift too, then if it is
audible, it would have to be at similar levels in both the live and
reproduced sound, and that would mean that Doppler shift is a natural part
of everything we hear, so in a speaker, it could not be considered
distortion at all!


By Occam's Rasor, either our hearing mechanism has built in
compensation,
so Doppler distortion doesn't matter, or the math is wrong and it needs
to
be revised. That isn't to say that it doesn't happen in the
piston-in-an-infinite-tube model, it just means that the speaker/room
model
is a totally different animal.

Sure it is but it should get signifigantly worse with a
driver in an enclosure rather than in a tube because of the
large excursions demanded at the low frequencies to couple
anything from an enclosed speaker to a room. You can get
real high SPL low frequencies in a tube without much
excursion, which is what causes it, but not so with an
enclosed speaker in a room. You have to push a whole lot of
LF air up close to get what's in the signal to reach you at
any distance.


It all depends on if the virtual sound source is the rest point of the
speaker cone, or if the instantaneous virtual sound source is the surface of
the cone at any given point in time. If it is the rest point of the cone,
then no Doppler shift occurs because that point is stationary relative to
the listener. If it is the surface of the cone at any given point, then
Doppler shift does occur in a speaker, but it also occurs in a guitar string
when it is plucked, in a drum head when it is struck, in a trumpet when a
note is being played, in everything that makes a sound by vibration (meaning
everything!)

Dunno what you mean by built in compensation nor why there
would be anything like that. It's not the kind of thing
evolution would have devoted much energy to. There weren't
many broadband sound sources to work with even if it had
been deemed important for some reason.

There are a huge number of natural broadband sound sources, thunder, wind,
the ocean, a babbling brook, a river, a herd of stampeding elephants, etc.
But I said "either there is some sort of built in compensation, or the
Doppler shift isn't audible." The "either" is the key word there, and I
think it is the latter, not the former.