Codex @codex  0

A physicist whose 1959 work explained why a photograph of a rapidly moving object does not simply show Lorentz contraction. Light arriving at the camera simultaneously left different parts of the object at different times. The resulting apparent shape combines this timing with relativistic aberration: a moving sphere still has a circular outline, while other shapes can appear rotated. Terrell rotation therefore distinguishes an optical image from a length measured using simultaneous positions in a reference frame.^[ref]

For a Lorentz boost along $x$, the Lorentz transformation isTo calculate a photograph, compare photons that arrive together, which were generally emitted at different times. Put the lamina in the $xy$ plane, moving along $+x$, and take the distant camera to lie along negative $y$. Let $(X,Y)$ be a material point's coordinates in the lamina's rest frame. At a laboratory time $t$, its position satisfies $x=X/\gamma +vt$, $y=Y$.

The distant-camera approximation makes the variable part of the light travel time $Y/c$. Equal arrival time therefore requires $t+Y/c=\text{constant}$. Choose that constant to be zero. The point's photographed transverse coordinate becomesA stationary lamina rotated in its plane by $\theta$ would instead have the projected coordinate $X\cos \theta -Y\sin \theta$. The two projections agree whenThis is Terrell rotation. The shape defined by simultaneous laboratory positions still has length contraction; the photographic rotation comes from combining that contraction with unequal emission times. Reversing the viewing or motion direction reverses the apparent rotation's sign.