The rough draft below shows the situation/Gedankenexperiment which I don´t understand.

The (clockwise) rotating object (e.g. a galaxy) emits photons from identical atoms (say H₂) at positions A and B, respectively. Photons A and B have the same wavelength/frequency.
There´s an observer at a constant distance from the rotating object who will see photon A blueshifted (relative to the emitted photon), and photon B redshifted.
There´s no doubt about this, we have seen photos of the Andromeda Galaxy which clearly show this effect.

But how can that be explained? There´s no relative motion between the photons (after emission) and the observer (in this Gedankenexperiment!).

I´m obviously missing something fundamental here.

Attached Images

redshift 4.JPG (39.5 KB, 25 views)

__________________
If everyone had even a basic grasp of scientific principles, this planet would be a better place (Phil Plait)

Die Lücke, die wir hinterlassen, ersetzt uns vollkommen [The gap we will leave behind will take our place entirely] (Carl Heinz Schroth)

1 + e^i*pi = 0

A is moving toward the observer at the moment the light is emitted, and B is moving away. That causes the blueshift and redshift respectively. The fact that the sources are moving in circles rather than straight lines does not change that immediately. Of course those shifts will change as the emitters come around in their orbits. Half an orbital period later A will be redshifted and B blueshifted.

But the frequency shift doesn´t happen at the moment of emission. That´s at least how I understand it. Both photons A and B have exactly the same wavelength when emitted at A and B, respectively.

__________________
If everyone had even a basic grasp of scientific principles, this planet would be a better place (Phil Plait)

Die Lücke, die wir hinterlassen, ersetzt uns vollkommen [The gap we will leave behind will take our place entirely] (Carl Heinz Schroth)

1 + e^i*pi = 0

Relative to the objects that emitted them. And those objects have different velocities relative to the detector.

When you look at spectra of stars you see absorption lines that are broadened by the speed of rotation of the star.

I think the trouble dhd40 may be having is in reconciling the particle vs. wave nature of light. If thought of as a wave, then it should be easy to picture how the Doppler effect works (and zillions of intro textbooks or popular books in physical science show how this happens using waves).

While I do not pretend to understand quantum field theory (QFT) at any level but shallow, I do understand that it is sometimes less useful to think of light as (classical) little particles traveling through space. First off, all quanta travel as wave packets. One can think of photons as what happens during an interaction between the quantized electromagnetic field and the quantum field of one or more electrons (usually) -- a localized exchange of energy and momentum. (Also -- the EM interaction between charged particles is mediated through exchanges of virtual photons.) An expanded discussion of the nature of photons can be found here.

The above description is probably a little squishy, but the gist is probably there.

Does that mean that the frequency shift happens already during emission? If so, how does the emitter know about the relative velocities between emitter and detector? And there are trillions of detectors with trillions of different relative velocities.

__________________
If everyone had even a basic grasp of scientific principles, this planet would be a better place (Phil Plait)

Die Lücke, die wir hinterlassen, ersetzt uns vollkommen [The gap we will leave behind will take our place entirely] (Carl Heinz Schroth)

1 + e^i*pi = 0

The emitter doesn't know or care what the detectors are doing. It just does its thing, which is to emit light that has a characteristic wavelength as measured by a detector which is stationary relative to the emitter. Detectors which are moving at different velocities get different values when they measure that wavelength.

You're putting the detector in the rest frame. Try looking at this from the emitter's point of view. The photon is emitted from point B at its rest frequency. The detector is moving away from the photon, so when the photon reaches the detector it takes a little bit longer for whole wave to enter the detector and the detector therefore sees a lower frequency (blueshifted) photon. A photon leaving from point A enters the detector more quickly due to the relative motion, and is read as having a higher frequency (redshifted).

That description may not be perfect, but Galileo will have to work right now. Adding length contraction and time dilation would probably make the above more accurate.

I'm sorry, but you appear to have it backward. In this line of thought, if a wave takes longer to enter a receding detector, it should appear to be a lower frequency and redshifted.

Ok so let us imagine that there are two detectors equal in distance from the emitter and directly opposite each other through the centre of the emmiter. Telemetry brings their data back to you. Now for detector 1 everything is as it was for the OP, photons from point A on the rotating object are blue shifted. Whilst those from point B are red shifted.

But for detector 2 photons from point A are red shifted and those from B are blue shifted.

Exactly the same amounts of shift but B swaps with A for detector 2.

So a photon leaving point A and propagating towards detector 1 has gained some energy due to the rotation from detector 1's perspective. Whilst a photon leaving point A and propagating to detector 2 has lost some energy from detector 2's perspective. However this does not mean that the rotation itself is losing energy due to the emmision of photons (this is just a relative effect) because for every higher energy photon emmitted there is a lower energy one to 'compensate' in a sense. It is epecially interesting to put the detectors on the same side but have a mirror where detector 2 was.

__________________
"You can't talk to a brick wall but you can do Graffiti"

By the way all these effects could be demonstrated with a loud speaker on a spinning wheel and a microphone as a detector.

__________________
"You can't talk to a brick wall but you can do Graffiti"

dhd40,

Here's an animation I made to explain a related (and more complex)
question about Doppler shifts. You can ignore the numbers on the right.
Pay attention to the bullets. Compare the progress of the bullets from
gunman number one and gunman number three. Each bullet can represent
a wavecrest in light emitted from a rotating galaxy. Gunman one is at the
center of the galaxy, not moving relative to us. Gunman three is moving
toward us, like matter at the top of your galaxy diagram.

http://www.freemars.org/jeff2/Doppler3.htm

Strangely enough, because of the strange fact that the speed of light is
the same for everyone, bullets fired by gunman number three are a better
analogy to light waves than are bullets fired by gunman four.

Bullets fired by gunmen one and three move at the same speed, but those
fired by gunman three are closer together because he moves forward
between shots, like the stars at the top of your rotating galaxy.

-- Jeff, in Minneapolis

__________________
http://www.FreeMars.org/jeff/

"I find astronomy very interesting, but I wouldn't if I thought we
were just going to sit here and look." -- "Van Rijn"

"The other planets? Well, they just happen to be there, but the
point of rockets is to explore them!" -- Kai Yeves

You know, that's what I wrote at first, and then I "fixed" it. I knew I should have just gone to bed. Thanks, Horblower.

I think what is bothering you is actually a very deep principle of relativity, that is often not well appreciated. The answer to "when does the frequency shift happen" is, it is not answerable in general. It is simply not a question that "absolute reality" knows the answer to, even though it sounds like it is a perfectly clear question. The only thing that is "absolutely real" is that a given measurement device receives a certain frequency from a given emitter. Just how that happened is a "story" that we tell, and the story can be told in many different ways, all equally valid if the measurables come out the same. In your case, you have been influenced by a particular version of how to tell the story, and because that version does not apply in general (as none of them do), you are confused as to why things don't work out.

Now, one solution some people take, especially in special relativity, is to make statements like "the answer is different for different observers". That's closer to the truth, but is not the truth, because even a given observer is not required to use any particular version of the story. What the "story" really relies on is an arbitrary choice of coordinate system, and if the same observer uses different coordinates, they get a different story about when and why the frequency shift occured. No observer is forced to use any particular coordinates, it is just that there is a common convention used in special relativity, and that convention is often erroneously interpreted as the "actual reality" for that observer. But there is no such actual reality, other than the outcome of the measurement itself. General relativity corrects this mistake, but it also brings in gravity, which makes it so complicated that the important message kind of gets lost.

Ken,

I think the actual truth is: “Nobody knows the answer yet.”

The same is true with light from the receding galaxies. Does the redshift take place when the light leaves the galaxies, or while it travels through space, or when it enters our galaxy, or as it reaches the earth?

The same applies to light the earth receives from a distant start that is fixed relative to the sun. When we are moving away from the star, during our annual rotation around the sun, we receive the star’s light as being redshifted, but when we move toward the star six months later, we receive its light as being blueshifted. This hints that the frequency shift takes place somewhere near the earth, perhaps inside our own solar system. But what causes it to shift, and where does the shift actually take place? I think the real answer is: “Nobody knows the answer yet”.

This is a simple Doppler effect regarding starlight, which Doppler himself predicted in the 1840s, so there is no need to confuse the issue with mysterious talk of “relativity” and assuming that the question is “not answerable”. The fact is, nobody yet knows how to answer it, but that doesn’t mean it’s not answerable, and I think it would be better to encourage new young astronomers to set a goal of trying to answer it in the future, rather than telling them to just forget about it because it’s “not answerable”.

dhd40. It's not just the photons either. Your object rotates in the background neutrino sea, in which we all live and observe. Viewed from your vantage point, the receding edge of a solid object runs into the neutrinos that strike it, effectively blue shifting them. The approaching edge runs away from them, effectively redshifting them. If originally they were relatively isotropic, they now are not. Cross-sections for both charged and neutral currents(the manner in which neutrinos slightly interact with normal baryonic matter) vary as the square of the frequency/energy. So, the effect the receding edge has on the object's "neutrino sea shadow" is more pronounced than the effect of the approaching edge. (Peltoniomi's Ultimate Neutrino Page). What does this do? From your distant vantage point the vector sum of the forces will be skewed slightly from the dead center of the rotating object...towards the receding edge. That means the orbiting satellite will precess it's perihelion ("periobjection") from considerations not including GR deformations of spacetime...curved space, but from considerations of straight-forward particle physics.That's not to say GR is not there, it's just not acting alone....not in our universe. pete

your object:http://www.bautforum.com/attachments...redshift-4.jpg



Ultimate Neutrino Page see:http://cupp.oulu.fi/neutrino/nd-cross.html

__________________
A third rate theory forbids.
A second rate theory explains after the fact.
A first rate theory predicts.
A. Lomonosov

Sam,

I disagree. The answer is known and is as Ken says. The Doppler shift
does not occur at a time or a place. It is a relationship between the
motion of the emitter and the motion of the receiver. Which is entirely
relative. Which is what relativity is about.

The basic question here is extremely simple so I am surprised that you
would assert that the answer isn't known.

-- Jeff, in Minneapolis

__________________
http://www.FreeMars.org/jeff/

"I find astronomy very interesting, but I wouldn't if I thought we
were just going to sit here and look." -- "Van Rijn"

"The other planets? Well, they just happen to be there, but the
point of rockets is to explore them!" -- Kai Yeves

At some level, that is the answer to all questions. But since it is clear that a better answer is of more value, what we really mean when we ask a question is, "what is the answer according to our current understanding of our reality". The answer to that question is the one I gave-- according to our current understanding, reality does not answer that question, period. Indeed, it is a central tenet of our best theory dealing with the question (relativity) that reality does not answer that question. Certainly, relativity could be wrong, that more or less goes without saying in science.
Even with the "simple" Doppler shift, the fact that light does not have a medium (apparently) is the reason that we cannot say when the shift occured. This holds true even for the simple Doppler shift, there's no need to include time dilation, the absence of an absolute frame is all that is required. Doppler himself would have assumed light had a "ponderable medium" (to use Einstein's words), so would have thought the question had an answer, but he would have been wrong (apparently). That he could get the right prediction, imagining light had a medium, does not mean that light does have a medium-- it was mere happenstance that his prediction did not require that tidbit of knowledge. Yet, today we do have that tidbit.
Then you misinterpret the key lessons of relativity. I would make an analogy with the uncertainty principle in quantum mechanics-- do you maintain that teaching the uncertainty principle is telling our new young astronomers to "just forget about" trying to get more information than that principle allows, or do you think it is an example of physics education? If you don't understand relativity yourself, that's no reason to not teach it to "young astronomers".

Well said, Ken. pete

__________________
A third rate theory forbids.
A second rate theory explains after the fact.
A first rate theory predicts.
A. Lomonosov

This is at least somewhat answerable, based on the excitation of certain molecules by the CMB as a function of redshift.

Actually, the question remains unanswered for each of those molecules. The point is, every observer, and every molecule, will see something, for sure, but when and why what they see became that way is never knowable, the result is only a function of the complete process, the null geodesic followed by the light. For example, if I perceive the molecules you describe as moving, or as being in some gravitational field, based on any particular coordinates I'm using to separate space and time in any arbitrary way, I may reason that it wasn't the light at all that had anything happen to it. My picture could as well be that the molecules had something happen to them, while the light was always just the same. There simply is no experiment that comes out A if something happened to the light, and B if it happened to the molecules, and that's why the question goes unanswered in a unique way-- we are free to answer it in any number of equally physically valid ways, as they can all be made entirely equivalent (indeed, the appearance of the word "equivalent" in the equivalence principle is no coincidence here-- equivalence is at the core of general relativity).

To make this more clear, imagine I am in a rocket moving toward the source of the CMB in some direction. Perhaps I'm moving so fast that the CMB looks just like the red light it was when it was first emitted, not the microwaves we see from Earth. Shall I say the light was redshifting all the while, then suddenly got blueshifted just as I observed it, or shall I say nothing changed about the light at all, it was just all those moving intervening molecules that thought the light was redshifting? Granted, the "comoving frame" description has plenty of merit as a convention, but it still is never anything but convention-- the question as posed is still not knowable, because convention is not truth.

(my bold)
But that´s exactly what I thought was not happening in my Gedankenexperiment, because the distance between the rotating object and the detector doesn´t change. Therefore, after the photon has been emitted it´s already moving in the detector´s frame (my view), isn´t it?

__________________
If everyone had even a basic grasp of scientific principles, this planet would be a better place (Phil Plait)

Die Lücke, die wir hinterlassen, ersetzt uns vollkommen [The gap we will leave behind will take our place entirely] (Carl Heinz Schroth)

1 + e^i*pi = 0

As I see from several answers now the conclusion seems to be that we simply don´t know where and when the frequency shift happens. This somehow surprises me because my home-made model, which put the shift-happening () into the detector (or the detector´s frame), worked so well for many years.

Now, that´s a little bit disappointing. On the other hand, the outcome of measurements (of frequency shifts) seems to be very consistent. What I mean by this is it isn´t abitrary. Everybody who runs the same experiment will get the same answer.

Hmm, mother nature seems to know how to run things without telling us her secrets.

Thank you all for your valuable contributions

__________________
If everyone had even a basic grasp of scientific principles, this planet would be a better place (Phil Plait)

Die Lücke, die wir hinterlassen, ersetzt uns vollkommen [The gap we will leave behind will take our place entirely] (Carl Heinz Schroth)

1 + e^i*pi = 0

I cerytainly can't agree that the place at which the frequency shift occurs is unknown. It occurs at the detector. Or do I misunderstand the question?

Another thought experiment to illustrate. Two experiments.

1/ Galaxy as before. Beam of light from either limb. In setad of single observer, two at same distance, but ravelleing towrads and away from the galaxy, at the velocity of the rim. 'Towards' at the receding limb.
Both will 'see' the light at the same wavelength.
If they stop, relative to the galaxy, they will both see equal frequency shift, in opposite directions.

2/ 'Observer' at galaxy rim.
'Sees' light from approaching atom, notes blue shift.
Turns quickly as the atom goes by, 'sees' red shift, equal amount from same atom, now receding.
But if the observer is in the galaxy , roattaing with the atom - no shift.

Wherever you place the 'observer', whatever their relative velocity to the radiator, the frequency shift occurs at the observer.

John

Not so, and neither of your examples establish your claim. You give two examples that show the motion of the detector affects the result. How does that show when and where the shift occured? Motion is relative, ergo. "relativity". You could just as easily alter the galaxy, and you'll also get a difference in the shift. Does that prove the shift happens at the galaxy?

The rotating object isn't emitting the photons in question, small parts of the object, which have a relative radial velocity to the detector, are emitting these photons. You cited a galaxy as the rotating object in your OP. A galaxy is made up of many individual stars, most of which will have a relative velocity when compared to the detector. The photons come from the stars, not just from the center of the galaxy.

This is why I am not a cosmologist.

I agree at one level it's kind of frustrating, but at another, it's kind of liberating. To me, it's a bit like learning that arithmetic can be done using other bases than base 10, and it still works and still gives equivalent results. We use base 10 so much that we begin to think numbers themselves have to work that way, but numbers are something deeper that transcends the base used to manipulate them. Apparently, our local concepts of space and time are like that too, for observers in relative motion at the same place and time, and this retreat from uniqueness into equivalence introduces an ambiguity in global extensions of our local concepts of space and time. The conventions we use to calculate using those extensions are just conventions like a choice of base. We can't really manipulate and communicate arithmetic quantities without choosing a base, but that doesn't mean the entities we are manipulating give a hoot what base we choose. We have to learn to separate what nature is doing from what we are doing, it's an important lesson that permeates through all of science.

Ken,
My point was that I compared the detector/observer in motion and at rest, relative to the emitter. Would you accept that EMR from hydrogen atoms (as suggested by the PO) is determined by the electron energy, and so is at a constant frequency? It does so regardless of the velocity of the detector, which will find it is as advertised if there is no relative motion. So it is irrelevant if it is the galaxy or the detector that moves. Indeed, ignoring all other objects, there is no way of knowing which it is that is moving.

The light leaves the emitter at a set frequency - it has not changed until it hits the detector, and any change depends on the relative velocity between emitter and detector. As Lord Meercat says, "!Simples!"

John

Both its energy and its frequency depend on the reference frame. But yes, in the frame of the emitting atom, and when it is at the location of that atom (say, as it is emitted), we do know its energy and frequency unambiguously. That is the last place and time for which that statement is true, until we absorb that photon at the location of some other atom and in the frame of that atom, and what happened in between is unknowable and unspecifiable in any kind of absolute way. The reason is because the reference frames in between are completely arbitrary, the choices we make will control all the outcomes, we will be doing it ourselves (just as you are). Hence, we cannot state unambiguously when or where the frequency changed, we control the answer ourselves.

In special relativity, we can say why it changed, which is the "relative motion," but in general relativity, we cannot even say why unambiguously, because there is not even an unambiguous concept of nonlocal relative motion. It's all dependent on the coordinate choice, which introduces an inescapable arbitrariness that is only resolved by restricting to results that are invariants of the theory.
Correct, and that is why your claim above is unsupportable.
The first conclusion is not supported by anything you've said, and even the second conclusion is not generally true, it is only supportable in the limited viewpoint of special relativity. Personally, I don't think we should even teach special relativity, because it leads so inevitably to misconceptions like that second conclusion, but that's not the key part of what I'm saying, because what I said above is true even in special relativity-- we have no way of knowing when the frequency change occured, reality does not arbitrate the point.

I can prove my assertion quite easily. If you believe that reality does arbitrate when the frequency shift occurs if atom A emits a photon and atom B absorbs it at a different frequency, then tell me an experiment that comes out X if the frequency shift occured only at the point of absorption, and not X if it occured in some other way. My guess is, you would try to say something like, take atoms stationary with respect to A and put them all along the path from A to B. They'd all see the photon as being at the same frequency as A emitted it. But I'll just ask, why those atoms? Why not a chain of atoms stationary with respect to B? Those atoms would "prove" the opposite assertion, that the frequency shift happened immediately upon emission. The real answer is, neither is a supportable assertion, it is unknowable because "it takes two to tango"-- the invariant frequency shift is set by the complete process only.