Lorentz Transformation

Why the Lorentz transformation matters

• GPS satellite corrections. GPS satellites move at ~3.9 km/s relative to the ground, giving γ − 1 ≈ 8.3 × 10⁻¹¹. Their clocks run slow by ~7 µs per day from this kinematic effect (combined with a +45 µs gravitational redshift gain, the net is +38 µs/day). Without applying the Lorentz transformation to satellite signals, GPS positions would drift by ~10 km per day within hours, rendering navigation useless.
• Particle accelerators. The Large Hadron Collider accelerates protons to 99.9999991% of c, where γ ≈ 7,460. Their lifetime in the lab frame is dilated by γ, beam energies are boosted by γ relative to rest mass, and decay products are forward-collimated into narrow cones by relativistic aberration — all direct Lorentz-transformation consequences. Accelerator design is impossible without this calculus.
• Cosmic ray muons reaching Earth. Muons created at ~15 km altitude have rest-frame lifetime 2.2 µs, classically allowing only ~660 m of travel before decay. They arrive at sea level in detectable numbers because their lab-frame lifetime is γ × 2.2 µs (with γ ≈ 30 for typical cosmic-ray muons), or equivalently because the 15 km atmosphere is length-contracted to ~500 m in the muon's frame. Either viewpoint, derived directly from the Lorentz transformation, agrees with the observed flux.
• Synchrotron radiation pattern. An electron in a circular accelerator emits in all directions in its instantaneous rest frame, but Lorentz-boosted into the lab frame this isotropic pattern beams forward into a cone of half-angle ~1/γ. For a 1 GeV electron (γ ≈ 2,000), the radiation cone is ~0.5 mrad — narrow enough to be used as a high-brightness X-ray source in storage rings.
• Maxwell's equations covariance. Lorentz first derived his transformation as the change of variables that leaves Maxwell's equations form-invariant. The Lorentz transformation is what makes electromagnetism consistent across moving frames; the magnetic field seen by a moving observer is partly the electric field of a stationary observer transformed by Lorentz boost. The unification of electricity and magnetism into a single antisymmetric field tensor F_μν only works in this geometry.
• Quantum field theory foundation. All relativistic quantum field theories — quantum electrodynamics, the Standard Model — are built on Lorentz invariance. Particle states are classified by representations of the Lorentz group (scalars, spinors, vectors, tensors). The Dirac equation, the Klein-Gordon equation, and Maxwell's equations are all manifestly Lorentz-covariant. Any observable Lorentz violation would require fundamental rewriting of physics.
• Astrophysical jets. Active galactic nuclei and gamma-ray bursts produce jets at Lorentz factors γ ~ 10–1000. The boosted emission appears as superluminal motion (apparent transverse velocity exceeds c) and dramatic Doppler boosting of the observed luminosity by factors of γ³ to γ⁴ — both quantitative tests of the Lorentz transformation on cosmic scales.
• Relativistic kinematics. Pion decay, beta decay, and any particle reaction's kinematic constraints follow from Lorentz-invariant 4-momentum conservation. The Mandelstam variables s, t, u used to label scattering processes are Lorentz invariants built from initial and final momenta — they're the same in any frame, simplifying calculations enormously.

The transformation in detail

• Forward and inverse. Forward boost from S to S' moving at +v along x: t' = γ(t − vx/c²), x' = γ(x − vt), y' = y, z' = z. Inverse boost: t = γ(t' + vx'/c²), x = γ(x' + vt') — same form with v → −v, exactly as relative-motion symmetry demands. Composing the two gives the identity, confirming consistency.
• Matrix form. In units where c = 1 and writing β = v/c, the transformation acts on the 4-vector (ct, x, y, z) as a 4×4 matrix Λ with Λ⁰⁰ = γ, Λ⁰¹ = Λ¹⁰ = −γβ, Λ¹¹ = γ, others diagonal 1. det Λ = 1 (proper transformation), and Λ leaves the Minkowski metric η = diag(−1, 1, 1, 1) invariant: Λᵀ η Λ = η. This is why Lorentz boosts are sometimes called pseudo-rotations.
• Rapidity parameterization. Define rapidity φ via tanh φ = β. Then γ = cosh φ and γβ = sinh φ. The boost matrix becomes a hyperbolic rotation: t' = t cosh φ − x sinh φ, x' = x cosh φ − t sinh φ. Successive boosts along the same direction add rapidities φ₁ + φ₂, just like rotation angles. This makes the velocity addition formula cleaner: tanh(φ₁ + φ₂) = (tanh φ₁ + tanh φ₂)/(1 + tanh φ₁ tanh φ₂).
• Numerical γ values. v = 0.1c → γ = 1.005 (0.5% effect); v = 0.5c → γ = 1.155 (15.5% effect); v = 0.866c → γ = 2 exactly; v = 0.9c → γ = 2.29; v = 0.99c → γ = 7.09; v = 0.999c → γ = 22.37; v = 0.99999c → γ = 224. Each additional 9 in 0.99…9c roughly multiplies γ by √10.
• Length contraction. A rod at rest in S with endpoints x_A and x_B (length L = x_B − x_A) is observed in S' at simultaneous time t'. Solve x = γ(x' + vt') for x_A and x_B at the same t', subtract: L' = L/γ. The rod is shorter in the frame where it's moving. Note: the contraction is along the direction of motion; perpendicular dimensions are unchanged.
• Time dilation. A clock at rest in S' ticks at proper time intervals Δτ. Its position in S' is x' = const, so successive ticks have x' equal. Plug into t = γ(t' + vx'/c²): Δt = γΔτ. The lab observes longer intervals between the moving clock's ticks. The proper time τ is the invariant — it's what the clock's own rest-frame measures.
• Velocity addition. Object has velocity u in S; S' moves at v relative to S. Differentiate the transformation and divide: u' = dx'/dt' = (u − v)/(1 − uv/c²). Verify: if u = c, u' = (c − v)/(1 − v/c) = c. Light's speed is c in every frame.
• Doppler effect. A photon with frequency ω₀ in the source frame is observed at ω = ω₀ √[(1 − β)/(1 + β)] for a receding source (longitudinal redshift). Transverse Doppler effect: ω = ω₀/γ purely from time dilation, with no classical analog. Both follow from Lorentz-transforming the photon's 4-momentum.

How to derive it

• Postulate 1: relativity. The laws of physics take the same form in all inertial frames. So the transformation between frames must be linear (preserving inertial trajectories) and reversible.
• Postulate 2: light speed constant. A light pulse satisfies x² + y² + z² = c²t² in any inertial frame. Both frames must agree that this defines the same set of events.
• Linearity. Most general linear transformation: t' = At + Bx, x' = Ct + Dx (and y' = y, z' = z by isotropy). Four unknowns A, B, C, D as functions of v.
• Origin condition. The origin of S' (x' = 0) moves at velocity v in S, so x = vt when x' = 0, giving D = −Cv/v… simpler: x' = 0 implies Ct + Dx = 0, so x/t = −D/C = v, hence C = −Dv. Substitute back: x' = D(x − vt).
• Light-cone invariance. Plug a light ray x = ct into the transformation: x' = D(c − v)t, t' = (A + Bc)t. Demand x'/t' = c: D(c − v) = c(A + Bc). Plug x = −ct: similar gives D(c + v) = c(A − Bc). Solve: A = D and B = −Dv/c².
• Symmetry condition. The inverse transformation has the same form with v → −v. This requires γ(v) γ(−v)(1 − v²/c²) = 1. Combined with γ(v) = γ(−v) (no preferred direction), gives γ² (1 − v²/c²) = 1, so γ = 1/√(1 − v²/c²).
• Final form. D = γ. Therefore x' = γ(x − vt), t' = γ(t − vx/c²). The full Lorentz transformation, derived from two postulates plus linearity and isotropy.

Common misconceptions

• "All frames see contraction in all directions." No — only along the direction of relative motion. Perpendicular dimensions are invariant. A meter stick moving along x is shorter in x but unchanged in y and z. This anisotropy is essential: if all directions contracted, the speed of light couldn't be isotropic in both frames.
• "Moving clocks tick slower from any frame, so one of them is really slow." No. The relation is symmetric: each frame sees the other's clocks tick slow. There's no contradiction because simultaneity differs between frames — the two clocks are never compared at the same place at two different times in a frame-independent way. The twin paradox is resolved by acceleration: the traveling twin is non-inertial, breaking the symmetry.
• "γ multiplies energy directly." Energy and momentum each pick up a γ: E = γmc², p = γmv. The 4-momentum (E/c, p) Lorentz-transforms exactly like the 4-position. Total relativistic energy E = γmc² includes both rest energy mc² and kinetic energy (γ − 1)mc². For low v, (γ − 1)mc² ≈ ½mv², recovering classical kinetic energy.
• "Length contraction is a visual phenomenon." No — it's a measurement of simultaneous positions of the rod's endpoints, not what the eye sees. What the eye sees is also affected by light-travel-time differences (Penrose-Terrell rotation): a moving sphere appears rotated rather than flattened. Length contraction refers to the simultaneous-coordinate measurement, not the photographic image.
• "Lorentz transformations are just rotations." They are pseudo-rotations in spacetime that preserve the Minkowski metric, not the Euclidean metric. They mix space and time, but they're not ordinary rotations because of the relative minus sign in the metric. Spatial rotations are a separate subgroup; Lorentz transformations include both spatial rotations and boosts.
• "Speed-of-light constancy means light slows in glass." The constancy postulate refers to the speed of light in vacuum, which is the universal c. In media with refractive index n, the phase velocity is c/n and the group velocity is generally less — but the underlying photons still travel at c between scattering events. Cherenkov radiation occurs when a charged particle exceeds the medium's c/n, while still traveling slower than c.
• "Galileo was wrong." Galilean transformation is the v ≪ c limit of Lorentz, accurate to part-per-million for terrestrial speeds. It is still the right physics for cars, planes, baseballs, and most engineering. Lorentz becomes essential only at v approaching c, in particle physics and astrophysics. Galileo's relativity principle (laws same in all inertial frames) is preserved exactly; only the transformation itself was generalized.
• "You can use Lorentz across accelerating frames." Lorentz transformations relate inertial frames. Accelerating frames need general relativity or a sequence of instantaneously-comoving inertial frames. For uniformly accelerated frames, Rindler coordinates are the appropriate generalization, and they exhibit Unruh radiation.