# Why do we have red shift and blue shift, and not only one?

1. May 27, 2008

### AntigenX

I hardly understand anything about relativity (as would be evident from my threads and posts), and thus, this dumb question.

If there is only length contraction, and not length expansion, why do we have two shifts, namely redshift and blue shift? should not only there be only blue shift due to length contraction? Red shift would imply length expansion, I suppose!

2. May 27, 2008

### Staff: Mentor

Before you start another wild goose chase (:tongue:), do you understand the ordinary Doppler shift for sound? After all, there's no length contraction or time dilation at all there, yet there are "two" shifts.

Study this: Doppler Effect

3. May 27, 2008

### AntigenX

Served me better .

My guess was (in this case), that the red/blue shift was due to apparent (wave)length contraction/expansion when an observer is in relative motion with the source of light.

4. May 27, 2008

### Staff: Mentor

The relativistic Doppler is nothing more than the classical Doppler (e.g. for sound waves) with time dilation included. The time dilation part only goes one way (red) but the classical Doppler goes both ways (red and blue). The reason that you still get red and blue shifts overall is that the classical Doppler effect is HUGE compared to the time dilation effect, even for relativistic velocities.

5. May 29, 2008

### calis

Nice this have been broaght up...

I agree, that light shifting is not a case of relativity whatsoever... but what I cannot get... If a visible light is just a small part of all wavelength... then how come that ultraviolet rays spred by an apparting star doesn't become visible as a violet light (after all ultraviolet ray frequency also get longer and reaches the visible light value...??? and why red light doesnot go out of the visible light area and doesn't became a FM radio wave... red whift shouldnt accure at all. just we see an apparting star in different colour

Last edited: May 29, 2008
6. May 29, 2008

### Staff: Mentor

This does happen. In fact, the redshift is the way that the velocity of distant stars or galaxies is measured.

7. May 29, 2008

### gggarnet

When mosquito flies near you,you can hear sharp voice and when mosquito flies away,the sound of mosquito becomes not sharp as before.

8. May 29, 2008

### calis

To DaleSpam
you helped me on the other thread. I hope you will help on this one too

you dind't get me right.
I know about the dopler effect - actually it is called hubble law - that by determining red shift we measure the speed of departuring thus we can measure current distance

what concernes me is that red shift shouldnt exist

since the visible light ends with violet (high freqency) and red(low frequency) than I undersnatd... if a low frequency light becomes lower it gets out of the visible area.... and if a too high frequence gets a little lower than it enters the visible light area.

So I understand it should be balanced equaly

9. May 29, 2008

### Staff: Mentor

If the source of "light" were completely featureless, then you wouldn't be able to observe any shift. But it's not. To measure a shift, one looks for specific spectral features (for example, hydrogen absorption lines) and measures the shift in wavelength of those features. This might help: Wiki: Redshift

10. May 29, 2008

### Staff: Mentor

I don't understand your comments. If the frequency of the electromagnetic waves goes down it is called redshift regardless of whether or not it goes in or out of the visible range.

11. May 30, 2008

### yuiop

Hi Calis,

Doc Al is right. All elements give off a unique signiture spectrum that is like a bar code. You can see the emission spectrums for iron and hydrogen here http://en.wikipedia.org/wiki/Emission_spectrum

It is easy to see that the signiture for iron is very distinct from hydrogen. Most stars are mostly hydrogen (at the surface anyway) and you would expect to see the barcode for hydrogen. If the basic signiture code for hydrogen is there, but shifted to the left or the right, you can calculate the degree of redshift. This is probably a gross over simplification, but I hope you get the general idea.

12. Aug 22, 2010

### dancing queen

hi,

where can i find a list of the stars
--------------------------------
1-shifting to red (getting away from us)
2-shifting to blue (getting closer to us)

I'm particularly interested in
--------------------------
1-Orion constellation

13. Aug 23, 2010

### Passionflower

I am not sure how you wish to support this statement DaleSpam.

The ratio between the relativistic Doppler shift of an approaching and retreating object with the same speed is exactly 1. So how could you argue that the relativistic Doppler shift includes a time dilation part that only goes one way but is small compared to the classical Doppler shift?

That relativistic time dilation plays no role in the relativistic Doppler factor is clear after we consider the relationship between the square of the relativistic Doppler factor and the change in radar roundtrip time for increasing or decreasing distances, the clock speed of the ship containing the mirror plays absolutely no role in this.

Last edited: Aug 23, 2010
14. Aug 23, 2010

### Staff: Mentor

The relativistic Doppler shift formula is:

$$\frac{f_r}{f_e}=\sqrt{\frac{c-v}{c+v}}=\frac{c}{c+v} \sqrt{1-\frac{v^2}{c^2}}=\frac{k}{\gamma }$$

Where γ is the relativistic time dilation factor and k is the classical Doppler shift. In general γ is much closer to 1 than k is, especially for v<<c.

Last edited: Aug 23, 2010
15. Aug 23, 2010

### JesseM

Suppose something is moving away from you at 0.6c, and sending out light signals once every 20 seconds according to its own clock. Then if there is no time dilation, then it sends out signals once every 20 seconds in your frame too, and during this time it has moved a distance of 0.6*20 = 12 light-seconds further away from you, so if we still assume the light signals travel at c in your frame, you'll only receive signals from the object once every 20+12=32 seconds. On the other hand, if it's moving towards you at 0.6c and sending out signals once every 20 seconds with no time dilation, then it will be 12 light-seconds closer to you each time it sends a signal, so the you'll receive signals from the object once every 20-12=8 seconds. And 1/32 and 1/8 are the numbers you get if you use the classical Doppler shift formula $$f = \frac{v + v_r}{v + v_s} f_0$$, assuming that the receiver is in the frame where light waves move at c (the rest frame of the luminiferous aether) so vr=0, the velocity of waves is v=c, the frequency f0 at which the waves are being sent is 1/20, and the velocity of the source is vs=-0.6c or vs=0.6c.

Now see what happens if we include relativistic time dilation. In this case, if the object is moving away at 0.6c and sending out signals once every 20 seconds according to its own clock, then its clock is slowed down by a factor of 1.25 in your frame, so in your frame it only sends out signals once every 20*1.25 = 25 seconds. So, if it's moving at 0.6c it will have moved further away from you by a distance of 25*0.6 = 15 light-seconds each time it sends a signal. So, you will only receive signals once every 25+15=40 seconds. On the other hand, if it's moving towards you at 0.6c, it'll be 15 light-seconds closer each time it sends a signal, so you'll receive a signal once every 25-15=10 seconds. 1/40 and 1/10 are the numbers you get if you use the relativistic Doppler shift formula $$f_{received} = f_{source}\sqrt{\frac{1 + v/c}{1 - v/c}}$$ with v=-0.6c or v=0.6c respectively, with fsource = 1/20. So, you can see that time dilation does figure into the relativistic Doppler shift formula, and that the relativistic blueshift of 1/10 is equal to 8/10 times the non-relativistic blueshift of 1/8, and likewise the relativistic redshift of 1/40 is equal to 8/10 times the non-relativistic redshift of 1/32.

16. Aug 23, 2010

### Passionflower

We all seem to agree on the formula, how to calculate the effect and on the results.

$$f_{received} = f_{source}\sqrt{\frac{1 + v}{1 - v}}$$

Approaching and retreating objects with speed 0.6 get a resp. shift factor of 2 and .5

Completely symmetrical!

Then I do not see how one can sensibly argue this case:

"The time dilation part only goes one way (red) but the classical Doppler goes both ways (red and blue). The reason that you still get red and blue shifts overall is that the classical Doppler effect is HUGE compared to the time dilation effect, even for relativistic velocities."

There is absolutely nothing in the formula that refers to time dilation!

17. Aug 23, 2010

### Staff: Mentor

Right, it is only when you express it as I did in post 14 that you can separate it out into a "time dilation" part and a "classical Doppler" part. The time dilation part is a function of v², so it is always red, but the classical part is a function of v, so it can be red or blue depending on the sign of v.

18. Aug 23, 2010

### JesseM

Yes, and do you agree with my analysis showing that the formula can be derived by considering both time dilation and the fact that the motion of the source causes it to be different distances from the receiver each time it sends a signal (with only the second being involved in the classical Doppler shift)? And do you agree that when you compare the relativistic Doppler shift to the classical Doppler shift, both the blueshifted frequency and the redshifted frequency in the relativistic formula are lower than the blueshifted and redshifted frequencies in the classical formula, with the difference between the two formulas being solely due to the time dilation factor? So the time dilation factor "only goes one way", shifting both blueshifted light and redshifted light to lower frequencies (more towards red) than what we'd predict if their was no time dilation?

To put it another way, if we assume in the classical formula that positive velocity indicates motion of the source towards the receiver, and we assume that the receiver is at rest in the frame where light waves move at c in all directions (the frame of the luminiferous aether), then the classical Doppler shift formula would be:

$$f_{received} = f_{source} \frac{c}{c - v}$$$$= f_{source} \frac{1}{1 - v/c} = f_{source} \frac{1}{\sqrt{1 - v/c}*\sqrt{1 - v/c}}$$

If we simply multiply the right side by the time dilation factor $$\sqrt{1 - v^2/c^2} = \sqrt{1 + v/c}*\sqrt{1 - v/c}$$, we get:

$$f_{received} = f_{source} \frac{\sqrt{1 + v/c}*\sqrt{1 - v/c}}{\sqrt{1 - v/c}*\sqrt{1 - v/c}} = f_{source} \frac{\sqrt{1 + v/c}}{\sqrt{1 - v/c}}$$

...which is the relativistic formula. So, the relativistic formula can just be obtained by multiplying the classical blueshift/redshift by the time dilation factor.

Last edited: Aug 23, 2010
19. Aug 23, 2010

### JesseM

If you look up the name of the star you're interested in on wikipedia, there'll be a chart at the upper right of the article with various bits of information, look at the "radial velocity" in the "astrometry" section and that'll tell you how fast it's approaching or receding from the Sun (a positive number means it's moving away from us, a negative number means it's moving towards us). For example, here is a list of the brightest stars in the Pleiades, you can click on any star name to go to an article about that star which should list its radial velocity--Alcyone has a radial velocity of 10.1 km/second, and Atlas has a radial velocity of 8.5 km/second. Likewise, here you can find a list of stars in the constellation of Orion, click any one to see its radial velocity. You can also enter the name of any star into the SIMBAD astronomical database to double-check wikipedia's numbers, and this will also give the redshift of the star--for example for Rigel, the line "Radial velocity / Redshift" gives result "V(km/s) 20.7 [0.9] / z(~) 0.000069 [0.000003]".

Incidentally, it's better to start a new thread rather than revive a thread from years ago like you did...also, it would be better to ask this type of question in the astronomy forum rather than the relativity forum.