Relativistic Headlight (Beaming) Effect

Bolt an ordinary lamp to the nose of a rocket. In the rocket's frame the lamp is boring: it glows evenly in every direction, the same flux out of the front as out of the back. Now watch that lamp from the launch pad as the rocket climbs to $\beta = 0.995$ . It is no longer a lamp. It is a searchlight. Half of everything it emits is packed into a forward cone $5.7^\circ$ across, and the light inside that cone is not merely rearranged — it is amplified, by about $1.6\times10^{5}$ , while the light going backwards is dimmed by the same factor. The beam is $2.5\times10^{10}$ times brighter than the lamp's own backward glow. Nothing was done to the lamp. No mirror, no reflector. The photons went where aberration sent them, carrying the energy and the arrival rate the Doppler factor gave them. This is relativistic beaming, and it is why half the quasars in the sky look like they have one jet when they almost certainly have two.

Key takeaways

Signature $(+,-,-,-)$ . Angle convention: $\theta$ is the lab-frame angle between the source's velocity and the photon's propagation direction. Then $\mathcal{D} = 1/[\gamma(1 - \beta\cos\theta)]$ .
Half the photons of an isotropic emitter arrive inside the cone $\cos\theta_{1/2} = \beta$ — exactly, at every speed. The famous $\theta_{1/2}\approx 1/\gamma$ is only the $\beta\to1$ asymptote: at $\beta = 0.5$ the truth is $60.0^\circ$ and $1/\gamma$ says $49.6^\circ$ .
Aberration and beaming are one phenomenon. Aberration says where the photons go; $\mathcal{D}$ says what they carry — and the compression factor is $\mathcal{D}$ . See aq-sr-relativistic-aberration.
Forward amplification: $\mathcal{D}^4$ bolometrically (energy $\times$ rate $\div$ solid angle), $\mathcal{D}^3$ in specific intensity, $\mathcal{D}^{3+\alpha}$ in flux density for a discrete blob with $S'_{\nu'}\propto\nu'^{-\alpha}$ — and only $\mathcal{D}^{2+\alpha}$ for a steady jet.
Owned result: $\gamma(1-\beta) \le \mathcal{D} \le \gamma(1+\beta)$ , with $\mathcal{D}_{\min}\mathcal{D}_{\max} = 1$ exactly ( $\mathcal{D}_{\max} = e^{\varphi}$ , $\mathcal{D}_{\min} = e^{-\varphi}$ ). So an approaching jet outshines its identical receding twin by $[(1+\beta)/(1-\beta)]^{4} = e^{8\varphi}$ — at $\gamma = 10$ , that is $2.5\times10^{10}$ .
Key idea: a fast source does not glow, it aims. Boosting an isotropic emitter makes a searchlight whose width is set by aberration and whose brightness is set by four powers of the Doppler factor — and those are the space part and the time part of one Lorentz transformation.

The core insight

An isotropic source has no preferred direction. A moving isotropic source does — the direction it is moving — and relativity hands that direction to the photons twice over. Aberration sweeps rays forward, so the same photons occupy a smaller solid angle. The Doppler factor blueshifts each one and speeds up their arrival. Multiply and the forward power per solid angle goes as $\mathcal{D}^4$ . No force acts on the light: this is pure kinematics of a null four-vector, seen from a frame in which the source moves.

Where this lives today

Blazars are extreme because of $\mathcal{D}^4$ . A blazar is an active galaxy whose jet points within a few degrees of our line of sight — and "a few degrees" is not a coincidence, it is $\arccos\beta$ . Inside that cone $\mathcal{D}\sim2\gamma$ , so a source of utterly ordinary intrinsic luminosity is amplified by $\mathcal{D}^4\sim10^5$ and lands in the gamma-ray catalogues. Blazars dominate the extragalactic gamma-ray sky not because they are the most powerful objects but because they are the ones aimed at us: the observed population is a Doppler-selected sample.

Apparent superluminal motion is the same factor in a different hat. A blob at speed $\beta$ and small angle $\theta$ appears to cross the sky at $\beta_{\text{app}} = \beta\sin\theta/(1-\beta\cos\theta)$ , which exceeds $1$ over a wide range of angles and peaks at $\gamma\beta$ — nearly $10$ for a $\gamma = 10$ jet. Nothing exceeds $c$ : the blob nearly keeps pace with its own light, so photons emitted over a long interval arrive bunched into a short one. The same $(1-\beta\cos\theta)$ sits in both denominators.

Synchrotron radiation is beaming applied to one electron. An ultrarelativistic electron spiralling in a magnetic field radiates, in its instantaneous rest frame, the gentle dipole pattern of any accelerated charge. Boost it and the dipole becomes a searchlight of width $\sim1/\gamma$ sweeping past the observer once per orbit, so the observer sees not a smooth wave but a train of very brief pulses — and the Fourier transform of a brief pulse is a broad, hard spectrum. The hardness of synchrotron radiation is a direct consequence of the headlight effect, and it is why storage rings are built at all.

Gamma-ray bursts push it furthest. Bulk Lorentz factors of a few hundred give opening angles of a few tenths of a degree ( $1/\gamma = 0.19^\circ$ at $\gamma = 300$ ) and boosts $\mathcal{D}^4\sim(2\gamma)^4 = 1.3\times10^{11}$ . This is what resolves the compactness problem: the observed fluence would imply an opaque fireball if it were emitted isotropically, but beaming shrinks the true energy budget and the source is transparent after all.

The physical picture

Put the source at rest in $S'$ , with $S'$ moving along $+\hat{x}$ at $\beta c$ relative to the lab $S$ . In $S'$ the lamp is isotropic, so the plane $\theta' = 90^\circ$ — perpendicular to the motion in the source's own frame — cuts the emission exactly in half. That is a fact about the lamp alone; the lamp does not know it is moving. Now ask where the dividing plane lands in the lab. Aberration maps a ray at $\theta' = 90^\circ$ onto $\cos\theta = \beta$ . The plane has become a cone, and half of everything the lamp emits is inside it. That deserves saying carefully, because the textbooks usually do not: $\theta_{1/2} = \arccos\beta$ is exact. The famous $1/\gamma$ is what $\arccos\beta$ becomes as $\beta\to1$ , and it is a poor approximation at moderate speed. At $\beta = 0.5$ the exact half-angle is $60.0^\circ$ ; $1/\gamma$ claims $49.6^\circ$ , $21\%$ too small. Only by $\beta = 0.99$ do they agree to better than a percent ( $8.110^\circ$ against $8.083^\circ$ ). Collimation alone would be a modest effect: squeezing photons into a cone of solid angle $\sim\pi\theta_{1/2}^2$ raises the intensity by roughly $\gamma^2$ — nowhere near enough to erase a counter-jet. The rest comes from the fact that a photon is not just a direction. It is a frequency and an arrival time, and both transform too.

Figure

The dividing plane becomes a cone. In the source's own frame (left) an isotropic lamp sends exactly half its photons forward of the plane θ′ = 90°. Aberration maps that plane onto the lab-frame cone cos θ = β (right) — EXACTLY, at every speed. The half-angle is arccos β, not 1/γ: at β = 0.5 the true value is 60.0° while 1/γ claims 49.6°, a 21% error. Only above β ≈ 0.99 does the 1/γ asymptote earn its keep

The core equations

\mathcal{D} \;\equiv\; \frac{\omega_{\text{obs}}}{\omega_{\text{src}}} \;=\; \frac{1}{\gamma\,(1 - \beta\cos\theta)}, \qquad \theta = \text{LAB angle between } \mathbf{v} \text{ and the photon's propagation direction} \qquad (1)

where: $\beta = v/c$ , $\gamma = (1-\beta^2)^{-1/2}$ . Physical reading: $\theta = 0$ is a photon flying along the source's velocity — straight at an observer the source is chasing. $\theta = \pi$ is a photon flying backwards from a source running away. The formula is derived from the invariant $k_\mu u^\mu$ in aq-sr-doppler-four-momentum; it is used, not re-derived, here.

\cos\theta = \frac{\cos\theta' + \beta}{1 + \beta\cos\theta'}, \qquad \sin\theta = \frac{\sin\theta'}{\gamma\,(1 + \beta\cos\theta')}, \qquad\text{and}\qquad \gamma(1 + \beta\cos\theta') = \mathcal{D} \qquad (2)

where: $\theta'$ is the emission angle in the source's rest frame. Physical reading: the last identity is the hinge of the article. The factor that compresses transverse angles is the Doppler factor. Aberration and Doppler boosting are not two conspiring effects — they are one number appearing in the space part and the time part of one four-vector. (Aberration itself is owned by aq-sr-relativistic-aberration.)

\boxed{\;\cos\theta_{1/2} = \beta \qquad\Longrightarrow\qquad \theta_{1/2} = \arccos\beta \;\xrightarrow{\ \beta\to1\ }\; \frac{1}{\gamma}\;} \qquad (3)

where: $\theta_{1/2}$ is the lab half-angle of the cone holding half of an isotropic emitter's photons. Physical reading: the arrow is an asymptote, not an equality.

I_\nu = \mathcal{D}^3 I'_{\nu'}, \qquad \frac{\mathrm{d}P}{\mathrm{d}\Omega} = \mathcal{D}^4 \frac{\mathrm{d}P'}{\mathrm{d}\Omega'}, \qquad S_\nu = \mathcal{D}^{3+\alpha} S'_\nu \ \ \text{for} \ \ S'_{\nu'}\propto\nu'^{-\alpha} \qquad (4)

where: $\alpha$ is the intrinsic spectral index. Physical reading: the exponent is not universal. $\mathcal{D}^4$ is the bolometric boost of a discrete moving source; a continuously fed jet — a steady tube, not a blob — loses one power, giving $S_\nu = \mathcal{D}^{2+\alpha}S'_\nu$ , because a stationary flow pattern does not enjoy the arrival-rate compression of a finite lump.

Where the beaming comes from

Derivation AThe half-angle of the forward cone, exactly

The claim is that a boosted isotropic emitter sends exactly half its photons into the lab cone $\cos\theta \ge \beta$ , and that this is exact at every speed rather than a large- $\gamma$ approximation. The algebra is short; the scoping matters more. The ubiquitous $\theta_{1/2}\approx1/\gamma$ is an ultrarelativistic asymptote which many sources print as though it held generally, and at $\beta = 0.5$ it is wrong by twenty-one percent. We derive the exact result first and take the limit only afterwards.

Step 1. Find the dividing surface in the source frame. Isotropy in $S'$ means the flux per solid angle is constant, so the fraction emitted into $\theta'\le\Theta'$ is $\tfrac12(1-\cos\Theta')$ . Setting that to $\tfrac12$ gives $\cos\Theta' = 0$ , i.e. $\Theta' = 90^\circ$ . Why: this is the only step that uses the source's physics, and it uses almost none — merely that the lamp is isotropic in its own rest frame.

Step 2. Aberrate the dividing surface. Put $\cos\theta' = 0$ into (2): $\cos\theta_{1/2} = (0+\beta)/(1+0) = \beta$ , so $\boxed{\theta_{1/2} = \arccos\beta}$ , exactly. Why: no photon is created or destroyed by a boost — only redirected. Half were forward of the plane; half are now inside the cone.

Step 3. Check that the map is monotone, so that "half" survives it. $\mathrm{d}(\cos\theta)/\mathrm{d}(\cos\theta') = (1-\beta^2)/(1+\beta\cos\theta')^2 > 0$ for all $|\beta|<1$ , so aberration is a strictly increasing bijection of $[-1,1]$ onto itself. Why: without monotonicity the median argument collapses. This is what guarantees exactly half, not approximately half.

Step 4. Only now take $\beta\to1$ . Write $1-\beta = (1-\beta^2)/(1+\beta) = 1/[\gamma^2(1+\beta)] \to 1/(2\gamma^2)$ ; since $1-\cos\theta_{1/2}\approx\theta_{1/2}^2/2$ , we get $\theta_{1/2}\approx\sqrt{2(1-\beta)}\approx 1/\gamma$ . Why: the $1/\gamma$ everybody quotes is this limit and nothing more. It needs the small-angle expansion of a cosine — exactly what fails at moderate $\beta$ .

Step 5. Quantify the failure. At $\beta = 0.5$ : $\arccos(0.5) = 60.000^\circ$ against $1/\gamma = 49.620^\circ$ , the asymptote $21\%$ low. At $\beta = 0.9$ : $25.842^\circ$ against $24.975^\circ$ ( $3.5\%$ ). At $\beta = 0.99$ : $8.1096^\circ$ against $8.0826^\circ$ ( $0.33\%$ ). At $\gamma = 10$ ( $\beta = 0.994987$ ): $5.7392^\circ$ against $5.7296^\circ$ ( $0.17\%$ ). Why: this is base defect D8, re-derived and quantified. The asymptote is excellent where jets live and useless where problem sets live, and the article printing only $1/\gamma$ has silently restricted itself to $\gamma\gtrsim5$ without saying so. *Verified in verify_aq-sr-relativistic-beaming.py: the aberration map is a monotone bijection; $\cos\theta_{1/2} = \beta$ symbolically; Monte-Carlo integration over an isotropic $S'$ distribution returns $0.500$ of the photons inside $\cos\theta\ge\beta$ for $\beta = 0.3, 0.6, 0.9, 0.99$ ; every degree value above is recomputed from the article text.*

Go deeper · Derivation B · 2 minWhy the boost is the fourth power of the Doppler factor

Derivation BWhy the boost is the fourth power of the Doppler factor

Collimation is the smaller part of the story. If aberration merely rearranged a fixed number of photons of fixed energy, the forward intensity would rise like the inverse solid angle, roughly $\gamma^2$ — impressive, but nowhere near enough to erase a counter-jet. The observed boost is far larger, because a photon carries an energy and an arrival time as well as a direction, and both are Doppler-shifted. This card counts the powers one at a time and then gets the exponent right for the source you actually have.

Step 1. Energy: one power. $E = \hbar\omega$ and $\omega_{\text{obs}} = \mathcal{D}\omega_{\text{src}}$ , so each photon arrives with its energy multiplied by $\mathcal{D}$ . Why: this is the Doppler shift proper — the only one of the three powers anybody remembers.

Step 2. Arrival rate: a second power. Photons emitted at proper rate $\mathrm{d}N/\mathrm{d}\tau$ arrive at rate $\mathcal{D}\,\mathrm{d}N/\mathrm{d}\tau$ . Why: the source is chasing its own photons, so two crests emitted a proper time $\Delta\tau$ apart arrive $\Delta\tau/\mathcal{D}$ apart. A clock tick is just another periodic signal — the wave and the clock are shifted by the same number. This is the step people forget.

Step 3. Solid angle: two more powers, upstairs. From (2), $\sin\theta = \sin\theta'/\mathcal{D}$ , and the longitudinal angle compresses at the same rate, so $\mathrm{d}\Omega = \mathrm{d}\Omega'/\mathcal{D}^2$ . Why: the same power flows through an aperture $\mathcal{D}^2$ times smaller, so power per unit solid angle rises by $\mathcal{D}^2$ .

Step 4. Assemble. \frac{\mathrm{d}P}{\mathrm{d}\Omega} = \frac{E\cdot(\mathrm{d}N/\mathrm{d}t)}{\mathrm{d}\Omega} = \frac{\mathcal{D}\cdot\mathcal{D}}{\mathcal{D}^{-2}}\,\frac{\mathrm{d}P'}{\mathrm{d}\Omega'} = \boxed{\mathcal{D}^4\,\frac{\mathrm{d}P'}{\mathrm{d}\Omega'}} Why: energy $\times$ rate $\div$ solid angle. Three relativistic effects, four powers of one number — exact, with no small-angle or large- $\gamma$ assumption anywhere.

Step 5. Cross-check against the invariant. $I_\nu/\nu^3$ is a Lorentz scalar (a photon phase-space density), so $I_\nu = \mathcal{D}^3 I'_{\nu'}$ at $\nu' = \nu/\mathcal{D}$ ; integrating with $\mathrm{d}\nu = \mathcal{D}\,\mathrm{d}\nu'$ gives $I = \mathcal{D}^4 I'$ — the same fourth power by a wholly different route. Why: two independent derivations agreeing is the check that the power-counting did not double-count or drop a factor.

Step 6. Get the exponent right. For a discrete blob with $S'_{\nu'}\propto\nu'^{-\alpha}$ , the flux density boosts as $\mathcal{D}^{3+\alpha}$ : four bolometric powers, minus one spent converting to a per-frequency quantity, plus $\alpha$ from evaluating the intrinsic power law at $\nu' = \nu/\mathcal{D}$ . For a steady, continuously fed jet the exponent drops to $2+\alpha$ , because a stationary flow pattern has no arrival-rate compression to gain. Why: at $\gamma = 10$ the difference between $\mathcal{D}^{3.7}$ and $\mathcal{D}^{2.7}$ is a factor of $\mathcal{D}\approx20$ . That is not a rounding error. *Verified in verify_aq-sr-relativistic-beaming.py: the identity $\gamma(1+\beta\cos\theta') = 1/[\gamma(1-\beta\cos\theta)] = \mathcal{D}$ holds identically under the aberration substitution; $\mathrm{d}\Omega/\mathrm{d}\Omega' = \mathcal{D}^{-2}$ by differentiating that map; and the two routes to $\mathcal{D}^4$ agree symbolically.*

Go deeper · Derivation C · 2 minThe range of the Doppler factor and the one-sided jet

Derivation CThe range of the Doppler factor and the one-sided jet

This is the article's owned structural result, and it repairs a defect that would have shipped a factor-of-two error in the faintest thing in the sky. The AllQFT base asserts that $\mathcal{D}$ is bounded below by $1/\gamma$ . It is not: $1/\gamma$ is the transverse value, an interior point of the range, and the true floor is smaller by a factor $(1+\beta)$ , tending to $1/(2\gamma)$ . Getting this right is what makes the one-sided-jet argument come out with the right exponent.

Step 1. Extremise. $\mathcal{D} = 1/[\gamma(1-\beta\cos\theta)]$ depends on $\theta$ only through $\cos\theta\in[-1,1]$ and is strictly increasing in it, so the extrema sit at the endpoints. Why: there is nothing to differentiate; a monotone function takes its extreme values where its argument does.

Step 2. Read off the ends. At $\theta = 0$ : $\mathcal{D}_{\max} = 1/[\gamma(1-\beta)] = \gamma(1+\beta)$ , using $\gamma^2(1-\beta^2) = 1$ . At $\theta = \pi$ : $\mathcal{D}_{\min} = 1/[\gamma(1+\beta)] = \gamma(1-\beta)$ . Hence \boxed{\;\gamma(1-\beta)\;\le\;\mathcal{D}\;\le\;\gamma(1+\beta)\;} Why: as $\beta\to1$ these tend to $1/(2\gamma)$ and $2\gamma$ . The base's floor $1/\gamma$ is a factor $(1+\beta)\to2$ too high — at $\beta = 0.9$ it claims $0.4359$ where the truth is $0.2294$ , a factor of $1.9$ .

Step 3. Locate $1/\gamma$ correctly. $\cos\theta = 0$ gives $\mathcal{D} = 1/\gamma$ : the transverse value, strictly inside the range for every $\beta>0$ . There is even a direction of no shift at all — $\mathcal{D} = 1$ needs $\cos\theta = (\gamma-1)/(\gamma\beta)$ , which at $\beta = 0.9$ is $\theta = 51.19^\circ$ . Why: a moving source has a redshifted hemisphere, a blueshifted cone, and one exact direction between them where its light arrives with its rest colour. The transverse direction is inside the redshifted part, well away from the floor.

Step 4. Notice the reciprocity. $\mathcal{D}_{\max}\mathcal{D}_{\min} = \gamma^2(1-\beta^2) = 1$ , exactly. In rapidity $\varphi = \operatorname{artanh}\beta$ : $\gamma(1\pm\beta) = \cosh\varphi\pm\sinh\varphi = e^{\pm\varphi}$ . Why: the longitudinal Doppler factors are the eigenvalues of the boost on the two null directions — recalled from aq-sr-lorentz-boosts-hyperbolic-rotations and aq-sr-the-light-cone. Forward and backward are exact reciprocals because a boost stretches one null ray by $e^{\varphi}$ and shrinks the other by $e^{-\varphi}$ .

Step 5. Build the one-sided jet. Two identical blobs are ejected back-to-back at speed $\beta$ , one at us ( $\theta = 0$ ), one away ( $\theta = \pi$ ). Equal intrinsic output, so R \;=\; \frac{\mathcal{D}_{\max}^4}{\mathcal{D}_{\min}^4} \;=\; \left(\frac{\gamma(1+\beta)}{\gamma(1-\beta)}\right)^{4} \;=\; \boxed{\left[\frac{1+\beta}{1-\beta}\right]^{4} = e^{8\varphi}} Why: because $\mathcal{D}_{\max}\mathcal{D}_{\min} = 1$ , the ratio is just $\mathcal{D}_{\max}^{8}$ — exponential in the rapidity, which is why it explodes. At $\beta = 0.6$ : $4^4 = 256$ . At $\beta = 0.9$ : $19^4 = 1.3032\times10^5$ . At $\gamma = 10$ : $(397.998)^4 = 2.509\times10^{10}$ .

Step 6. Draw the conclusion. A counter-jet suppressed by $2.5\times10^{10}$ is not faint; it is gone. Which is why the sources whose counter-jets we do see — M87's inner jet shows a jet-to-counter-jet ratio of only about $10$ to $15$ — must be slow, or nearly in the plane of the sky, or both. Why: run the argument backwards and the brightness ratio becomes a velocity meter. It is one of the very few handles astronomy has on a jet's bulk speed at all. *Verified in verify_aq-sr-relativistic-beaming.py: $\mathcal{D}$ monotone in $\cos\theta$ ; $\mathcal{D}_{\max}\mathcal{D}_{\min} = 1$ identically in $\beta$ ; $\gamma(1\pm\beta) = e^{\pm\operatorname{artanh}\beta}$ identically; $\mathcal{D}_{\min} = \gamma(1-\beta) < 1/\gamma$ for every $\beta>0$ (the D4 correction); the zero-shift angle $51.19^\circ$ at $\beta = 0.9$ ; and $R = [(1+\beta)/(1-\beta)]^4$ — the fourth power, asserted symbolically against the square.*

Figure

The full range of the Doppler factor at β = 0.9, against the LAB angle θ between the source velocity and the photon direction. 𝒟 runs from γ(1−β) = 0.2294 dead astern to γ(1+β) = 4.3589 dead ahead — and their product is exactly 1. The transverse value 1/γ = 0.4359 sits INSIDE the range, at θ = 90°: it is not the minimum, and the AllQFT base that says it is understates the faintest ray by a factor of 1.9. There is also one direction, θ = 51.19°, where a moving source shows its true colours: 𝒟 = 1 exactly

Figure

PUNCHLINE — the one-sided jet. Two identical blobs, ejected back-to-back at speed β. Plot the observed brightness RATIO of the approaching one to the receding one, against RAPIDITY φ = artanh β. Because 𝒟max = e^φ and 𝒟min = e^−φ, the ratio is R = e^(8φ) — a straight line on a log axis, 3.474 decades of brightness per unit of rapidity. At a typical blazar bulk Lorentz factor γ = 10 the counter-jet is fainter by 2.5 × 10^10. It is not dim. It is invisible

Limiting cases

Case	Result	Meaning
$\beta\to0$	$\mathcal{D}\to1$ , $\theta_{1/2}\to90^\circ$	no beaming: an isotropic lamp stays isotropic
$\beta\ll1$	$\mathcal{D}\approx1+\beta\cos\theta$	the classical first-order Doppler shift; no collimation to speak of
$\theta = 0$	$\mathcal{D} = \gamma(1+\beta) = e^{\varphi}\to2\gamma$	maximum blueshift, maximum boost
$\theta = \pi/2$	$\mathcal{D} = 1/\gamma$	pure time dilation — an interior point (see aq-sr-transverse-doppler-effect)
$\theta = \pi$	$\mathcal{D} = \gamma(1-\beta) = e^{-\varphi}\to1/(2\gamma)$	the true minimum, not $1/\gamma$
$\cos\theta = (\gamma-1)/(\gamma\beta)$	$\mathcal{D} = 1$	the one direction showing the source's rest colour
$\beta\to1$	$\theta_{1/2}\to1/\gamma$ , $R\to(2\gamma)^8 = 256\gamma^8$	the searchlight limit — but $\arccos\beta$ was exact all along
Continuous jet	$S_\nu = \mathcal{D}^{2+\alpha}S'_\nu$	one power less than a discrete blob

Common trapThe transverse value is not the minimum — and 1 over gamma is not the half-angle

Two independent errors, both live in the AllQFT base for this article, both involving $1/\gamma$ . D4: the range is $\gamma(1-\beta)\le\mathcal{D}\le\gamma(1+\beta)$ . The value $1/\gamma$ is what you get at $\theta = 90^\circ$ — an interior point. The floor is $\gamma(1-\beta)$ , smaller by a factor $(1+\beta)$ and tending to $1/(2\gamma)$ ; at $\beta = 0.9$ the base's floor is $0.4359$ and the truth is $0.2294$ . D8: the half-angle is $\arccos\beta$ , exactly; its $\beta\to1$ limit is $1/\gamma$ , and only its limit — printing $\theta_{1/2}\approx1/\gamma$ as a general result is $21\%$ wrong at $\beta = 0.5$ . Both mistakes share a cause: $1/\gamma$ shows up in three unrelated places in this subject, and it is easy to put it in a fourth where it does not belong.

Worked examples

1 · A blazar at gamma = 10 — how much of its counter-jet can we hope to see?

A quasar ejects two identical plasma blobs back-to-back along a jet axis pointing almost exactly at us. The bulk Lorentz factor is $\gamma = 10$ , squarely inside the observed blazar range of roughly $5$ to $40$ . Find the beam width, the two Doppler factors, and the brightness ratio.

Step 1 — identify. Two things are wanted: angular concentration (aberration) and brightness boost (four powers of $\mathcal{D}$ ). Both need $\beta$ : $\gamma = 10 \Rightarrow \beta = \sqrt{1 - 1/\gamma^2} = 0.9949874$ .

Step 2 — the beam width. $\theta_{1/2} = \arccos(0.9949874) = 0.100167\ \text{rad} = 5.7392^\circ$ . The $1/\gamma$ asymptote says $0.1\ \text{rad} = 5.7296^\circ$ ; here, at last, it is good ( $0.17\%$ ). Half of everything the blob emits lies inside a cone under six degrees wide.

Step 3 — the two Doppler factors. Approaching ( $\theta = 0$ ): $\mathcal{D}_{\max} = \gamma(1+\beta) = 19.9499$ . Receding ( $\theta = \pi$ ): $\mathcal{D}_{\min} = \gamma(1-\beta) = 0.0501256$ . Reciprocity check: $19.9499\times0.0501256 = 1.000000$ ✓. (Had we used the base's claimed floor $1/\gamma = 0.1$ , we would already be a factor of two adrift.)

Step 4 — apply. R = \left(\frac{\mathcal{D}_{\max}}{\mathcal{D}_{\min}}\right)^{4} = \left(\frac{1+\beta}{1-\beta}\right)^{4} = (397.998)^{4} = 2.509\times10^{10}, equivalently $\mathcal{D}_{\max}^{8} = 19.9499^{8}$ , or $e^{8\varphi}$ with $\varphi = \operatorname{artanh}(0.9949874) = 2.9932$ .

Step 5 — the boundary number. Let the approaching jet be an easily-imaged one-jansky source. The counter-jet then arrives at $4\times10^{-11}$ Jy, roughly four orders of magnitude below the microjansky noise of the deepest radio images ever made. We have not failed to detect it; detecting it is kinematically forbidden. Two identical jets — one visible, one erased, by geometry alone. Run the argument backwards and it becomes a velocity meter: where a counter-jet is seen (M87's inner jet shows a ratio of only about $10$ to $15$ ), the source cannot be both fast and aimed at us.

2 · A synchrotron blob at beta = 0.9 with a power-law spectrum

Same geometry at $\beta = 0.9$ , with intrinsic spectrum $S'_{\nu'}\propto\nu'^{-0.7}$ — a typical optically-thin synchrotron index. What is the observed flux-density ratio?

Step 1 — identify. Flux density, not bolometric power, and a discrete blob, so the exponent is $3+\alpha = 3.7$ .

Step 2 — apply. $\gamma(0.9) = 2.294157$ , so $\mathcal{D}_{\max} = 4.35890$ and $\mathcal{D}_{\min} = 0.229416$ , with ratio $(1+\beta)/(1-\beta) = 19$ exactly.

Step 3 — the three answers. Bolometric: $19^4 = 1.3032\times10^{5}$ . Flux density, discrete blob: $19^{3.7} = 5.388\times10^{4}$ . Flux density, continuously fed jet: $19^{2.7} = 2.836\times10^{3}$ — smaller by exactly one further factor of $19$ .

Step 4 — the boundary number. Three defensible exponents, three answers spanning $19^{1.3} = 45.96$ . Not one is wrong; they answer different questions about different sources. "How bright" is an incomplete question until you say what is moving and what you are measuring.

Step 5 — cross-check the collimation. $\arccos(0.9) = 25.842^\circ$ , against $1/\gamma = 24.975^\circ$ . Even at a respectable relativistic speed the asymptote is $3.5\%$ off — in the direction of making the beam look tighter than it is.

Physics of the portrait

Origins · 2 minRelativistic beaming and how this was discovered

Relativistic beaming

1898–1971 · Paris · Göttingen · Cambridge · Owens Valley

The mathematics arrived before the physics. Alfred-Marie Liénard (1898) and, independently, Emil Wiechert (1900) wrote down the retarded potentials of an arbitrarily moving charge, and there — in the denominator, where nobody expected a revolution — sat the factor $(1 - \boldsymbol{\beta}\cdot\hat{\mathbf{n}})$ . Cubed in the radiation field, it is the entire beaming effect: written down three years before relativity and seven before anyone knew what it meant. Einstein supplied the meaning in the last two sections of Zur Elektrodynamik bewegter Körper (1905): §7 derives aberration and the Doppler formula together from the transformation of a plane wave's phase, and §8 shows that the energy of a bundle of light transforms by exactly the same factor as its frequency — the observation from which every power of $\mathcal{D}$ in this article descends. Then the subject went quiet for sixty years, until the quasars woke it. In 1966 Martin Rees, a graduate student, pointed out in a two-page Nature letter that a relativistically expanding source seen at a small angle would appear to fly apart faster than light — an illusion produced by the very kinematics that beams it. Five years later Whitney and colleagues and Cohen and colleagues, using the new very-long-baseline interferometers, watched components of the quasar 3C 279 separate at an apparent three to four times $c$ . The theory was Rees's; the illusion was real; and beaming stopped being a footnote in electrodynamics and became the principal tool for weighing the speed of a jet.

History

1898

— Alfred-Marie Liénard derives the retarded potentials of a moving charge; the factor (1 − β·n̂) enters physics.

1900

— Emil Wiechert obtains the same potentials independently.

1905

— Einstein, in sections 7 and 8 of Zur Elektrodynamik bewegter Körper, derives aberration and the Doppler formula, and shows that the energy of a light bundle transforms exactly as its frequency.

1966

— Martin Rees predicts that relativistically expanding radio sources will appear to separate superluminally (Nature 211, 468).

1971

— Whitney and colleagues, and Cohen and colleagues, detect apparent superluminal motion in the quasar 3C 279 by very-long-baseline interferometry, at three to four times the speed of light.

Going further

Texts. Rybicki & Lightman, Radiative Processes in Astrophysics, §4.8 (the invariance of $I_\nu/\nu^3$ and the beaming exponents) · Rindler, Relativity: Special, General and Cosmological, 2nd ed., §4.4 · Jackson, Classical Electrodynamics, 3rd ed., §11.3 and ch. 14 · Lind & Blandford, Astrophys. J. 295, 358 (1985), on the blob-versus-jet exponents · Rees, Nature 211, 468 (1966).

Provenance. Signature $(+,-,-,-)$ ; $\theta$ is the lab-frame angle between the source velocity and the photon's propagation direction (band-locked). All numbers recomputed in verify_aq-sr-relativistic-beaming.py. Base defects repaired. D4: the base printed $\gamma^{-1}\le\mathcal{D}\le\gamma(1+\beta)$ ; the true floor is $\gamma(1-\beta)$ , a factor $(1+\beta)$ smaller ( $0.2294$ , not $0.4359$ , at $\beta = 0.9$ ) — and the base's own symbol gloss contradicted its own equation. D8: the base printed $\theta_{1/2}\approx1/\gamma$ as general; it is the $\beta\to1$ asymptote of the exact $\arccos\beta$ , $21\%$ wrong at $\beta = 0.5$ . Found here: the base's aberration card glosses $\gamma(1+\beta\cos\theta')$ as "the $1/\mathcal{D}'$ compression factor"; that quantity equals $\mathcal{D}$ , not its reciprocal — the mislabelling hides the central identity of the subject.

Refused. The lane specification gave the jet-to-counter-jet ratio as $[(1+\beta)/(1-\beta)]^{2}$ . It is wrong: $\mathcal{D}_{\text{app}}/\mathcal{D}_{\text{rec}} = (1+\beta)/(1-\beta)$ , so the fourth power of the ratio is $[(1+\beta)/(1-\beta)]^{4}$ . At $\gamma = 10$ the two differ by a factor of $361$ ( $2.51\times10^{10}$ against $1.58\times10^{5}$ ). This article prints the fourth power; the verify script asserts it symbolically.

Scope. Aberration is derived in aq-sr-relativistic-aberration; the Doppler factor in aq-sr-doppler-four-momentum; the transverse case in aq-sr-transverse-doppler-effect. The radiation dynamics — Larmor's formula, the synchrotron spectrum, the Liénard–Wiechert fields — are electrodynamics, and are not derived here.

Builds on. Relativistic Aberration of Light · Relativistic Doppler and the Photon Four-Momentum · Longitudinal Relativistic Doppler Effect. Leads to. Transverse Doppler Effect · Thomas–Wigner Rotation.

Is beaming just aberration, or is something else going on?

Aberration is exactly half of it — the half that decides where the photons go. It conserves photon number, and if that were all, the forward intensity would rise by the inverse solid-angle compression, about $\gamma^2$ . The other half is that each of those photons arrives blueshifted (one power of $\mathcal{D}$ ) and they arrive faster (a second). Multiply by the $\mathcal{D}^2$ of solid-angle compression and you have $\mathcal{D}^4$ . At $\gamma = 10$ that second $\mathcal{D}^2 = 398$ : calling beaming "just aberration" understates the forward brightness by a factor of four hundred.

Why is the exponent sometimes 3, sometimes 4, and sometimes 3 + α?

Because "brightness" is four different quantities. Specific intensity $I_\nu$ boosts as $\mathcal{D}^3$ , from the invariance of $I_\nu/\nu^3$ . Bolometric power per solid angle picks up one more power from $\mathrm{d}\nu = \mathcal{D}\,\mathrm{d}\nu'$ , giving $\mathcal{D}^4$ . Flux density of a discrete blob at fixed observed frequency is $\mathcal{D}^{3+\alpha}$ — three powers plus $\alpha$ from evaluating $\nu'^{-\alpha}$ at $\nu' = \nu/\mathcal{D}$ . Flux density of a steady, continuously fed jet is $\mathcal{D}^{2+\alpha}$ , one power less. Always ask: is the emitter a lump or a pipe, and am I measuring per-frequency or total?

If half the photons go forward, why is the forward brightness up by so much more than a factor of 2?

Because "half the photons" is an angular statement and the cone is tiny. At $\gamma = 10$ the cone of half-angle $5.739^\circ$ subtends $2\pi(1-\beta) = 0.0315\ \text{sr}$ — exactly a fraction $(1-\beta)/2 = 1/399$ of the sky. So half the photons occupy a quarter of a percent of the available directions: a number-density concentration of $1/(1-\beta) = 199.5$ before any Doppler shift at all. Then each on-axis photon is $\mathcal{D} = 19.95$ times more energetic and arrives $19.95$ times more often, and the on-axis power per solid angle is up by the full $\mathcal{D}^4 = 1.584\times10^{5}$ .

Is the half-angle really 1 over gamma, or is that an approximation?

An approximation — specifically, the $\beta\to1$ asymptote. The exact statement is $\cos\theta_{1/2} = \beta$ . At $\beta = 0.5$ that is $60.0^\circ$ while $1/\gamma$ claims $49.6^\circ$ : a $21\%$ error, in the direction of making the beam look tighter than it is. At $\beta = 0.9$ the error is $3.5\%$ ; at $\beta = 0.99$ , $0.33\%$ . For blazar and synchrotron work, where $\gamma\gtrsim5$ , the asymptote is entirely adequate and everyone uses it. For a problem at $\beta = 0.6$ , it is simply wrong.

Can beaming make something appear to move faster than light?

Apparently, yes — and it is the same algebra. A blob at speed $\beta$ and angle $\theta$ appears to traverse the sky at $\beta_{\text{app}} = \beta\sin\theta/(1-\beta\cos\theta)$ . That denominator is the one inside $\mathcal{D}$ , and it can be very small; maximised over $\theta$ it gives $\beta_{\text{app}}^{\max} = \gamma\beta$ , so a $\gamma = 10$ jet can appear to move at nearly ten times $c$ . Nothing exceeds $c$ : the blob nearly keeps pace with the light it emits, so a long journey arrives compressed into a short interval. Rees predicted it in 1966; 3C 279 showed it in 1971 at three to four times $c$ .

At what angle does a moving source show no Doppler shift at all?

Solving $\mathcal{D} = 1$ gives $\gamma(1-\beta\cos\theta) = 1$ , hence $\cos\theta = (\gamma-1)/(\gamma\beta)$ . This angle exists for every $\beta>0$ : it runs from $90^\circ$ as $\beta\to0$ down towards $0^\circ$ as $\beta\to1$ , and at $\beta = 0.9$ it is $51.19^\circ$ . It is the exact boundary between the blueshifted cone and the redshifted remainder — and it is not the transverse direction, where the light is already redshifted by $1/\gamma$ from time dilation alone. Good sanity check: if your Doppler formula has no zero-shift direction strictly inside the forward hemisphere, a sign is wrong.

Relativistic Headlight (Beaming) Effect

Where this lives today

The physical picture

The core equations

Where the beaming comes from

Limiting cases

Worked examples

1 · A blazar at gamma = 10 — how much of its counter-jet can we hope to see?

2 · A synchrotron blob at beta = 0.9 with a power-law spectrum

Physics of the portrait

Going further

Before this

Where this leads