Special relativity · Electrodynamics §7

Frequency and direction transform together
from the invariance of the phase.

How do the frequency and propagation direction of light transform between frames?

Read §7 of the 1905 relativity paper →

SR-09 · Doppler and aberration

Doppler principle and aberration

Ideal model, host calculation

The frequency and direction of light transform together because a plane wave's phase is invariant between frames. Observers in relative motion see the same light wave shifted in frequency and arriving from a different apparent angle.

Under a boost along x, a light ray at angle θ in K has frequency ratio ν'/ν = γ(1 - β cos θ) and apparent angle cos θ' = (cos θ - β)/(1 - β cos θ). Along the line of motion, the Doppler factor is √(1-β)/√(1+β) for receding and √(1+β)/√(1-β) for approaching rays.

Type a speed and an angle, or choose a named ray. Frequency and direction come from one wave-vector transform. The transverse case (θ = 90°) is the purely relativistic shift: the medium formulae give no change.

Set the observer and the ray
Relativistic Doppler and Aberration Vector DiagramStationary Frame KSource frame · ν = 500.0 THz · θ = 0.0°xyθ = 0.0°Moving Frame k (β = 0.600c)Observer frame · ν' = 250.0 THz · θ' = 0.0°x'y'v = 0.60cθ' = 0.0°
Doppler factor ν'/ν
0.500000
γ(1 - β cos θ)
Aberration cos θ'
1.000000
(cos θ - β)/(1 - β cos θ)

Accepted snapshot

Relativistic factors from one wave-vector transform, beside the two medium formulae that are right for sound.
Relativistic ν′/ν0.5
Lorentz factor γ1.25
Medium, moving observer0.4
Medium, moving source0.625
Receding line of sight0.5
Approaching line of sight2
Angle in k0 °

At θ = 90° the two medium formulae both give 1. The paper's factor is γ. That is the transverse Doppler shift, which has no classical counterpart. The medium formulae are not declared refuted: they are right for sound and agree with the paper to first order in v/c.

Not modeled: media and dispersion; sound in a medium; gravitational redshift; finite packets (SR-10); telescope optics and atmospheric refraction; photon picture; canal-ray apparatus beyond published values.

Worked case (readable without JavaScript)

Consider a light wave with frequency ν = 500 THz propagating at angle θ = 0° along the x-axis in the stationary system K, viewed by an observer moving along the x-axis at speed v = 0.6c (β = 0.6, γ = 1.25).

ν=νγ(1βcosθ)=ν1β1+β=0.5ν=250 THz\nu' = \nu\gamma(1 - \beta\cos\theta) = \nu\sqrt{\frac{1-\beta}{1+\beta}} = 0.5\nu = 250\text{ THz}

For a ray at right angles in the stationary system (θ = 90°):

ν=γν=1.25ν=625 THz,cosθ=β=0.6    θ126.87\nu' = \gamma\nu = 1.25\nu = 625\text{ THz},\qquad \cos\theta' = -\beta = -0.6\implies \theta' \approx 126.87^\circ

The phase φ = k·x − ωt is an invariant scalar that takes the exact same numerical value in both reference frames at every spacetime event.