The Doppler effect is the change in the observed frequency of a wave associated with motion of its source, its observer, or both. A familiar example is the higher pitch of an approaching siren and the lower pitch heard after it passes. The effect occurs with sound, light, radio waves, and other propagating disturbances. It connects measured frequencies with motion and therefore provides a foundation for velocity measurements across science and technology.
Its precise description depends on the physical system. Sound propagates through a material medium, so the source’s and observer’s velocities relative to that medium generally matter separately. For electromagnetic radiation in a vacuum, the relevant framework is special relativity, which combines propagation geometry with the different rates of moving clocks. Doppler measurements reveal stellar and galactic motions, track spacecraft, map winds, measure blood flow, and probe the motion of individual atoms. Interpreting them requires distinguishing a genuine frequency shift from changes caused by gravity, cosmic expansion, instrumental drift, or changes within the emitting material.
Discovery and experimental development
The effect is named after the Austrian physicist and mathematician Christian Doppler, who presented his proposal in Prague in 1842. His paper concerned the colored light of double stars and other stars. Doppler argued that motion between an observer and a wave source should change the rate at which successive wavefronts arrive, thereby changing the observed frequency. The underlying idea was correct, but his proposed explanation of contrasting stellar colors was not: stellar colors depend principally on surface temperature, and ordinary stellar velocities are far too small to produce the dramatic visible-color changes he envisaged.
Sound offered a more accessible test because ordinary terrestrial speeds are a substantial fraction of the speed of sound compared with their fraction of the speed of light. In 1845, the Dutch scientist Christophorus Buys Ballot tested Doppler’s prediction using musicians carried on a moving railway vehicle and listeners stationed beside the track. The observed changes in musical pitch supported the acoustic prediction. Such experiments made the phenomenon concrete without requiring the spectral resolution necessary for astronomical work.
In 1848, the French physicist Hippolyte Fizeau independently discussed the optical effect and emphasized the displacement of identifiable spectral lines. This was a decisive improvement over treating a star’s overall color as the observable. A spectral line supplies a sharply defined wavelength that can be compared with its laboratory position. For this reason, the optical phenomenon has sometimes been called the Doppler–Fizeau effect.
The growth of spectroscopy made astronomical velocity measurements possible. In 1868, William Huggins reported a line-of-sight motion for Sirius from the displacement of a spectral line. Early measurements were limited by weak stellar light, imperfect wavelength standards, and instrumental stability, but photographic and later electronic detectors greatly improved their precision. Doppler spectroscopy subsequently became a central method of astronomy, rather than merely a confirmation of a wave phenomenon.
The twentieth century extended the concept in two directions. Albert Einstein incorporated light-frequency transformations into special relativity in 1905, predicting effects that have no purely classical acoustic counterpart. Herbert Ives and G. R. Stilwell’s 1938 experiment with fast-moving hydrogen ions tested the relativistic contribution associated with time dilation. Meanwhile, radio engineering developed practical methods for measuring small frequency changes in reflected signals, establishing Doppler techniques as routine tools for navigation, remote sensing, and speed measurement.
Classical theory for sound and other waves in a medium
For sound, the simplest treatment assumes a homogeneous medium at rest, a constant wave speed , and a source emitting at frequency . During one emission period , a stationary source’s preceding wavefront travels a distance . The wavelength is therefore . A moving source emits its next wavefront from a different position, changing the separation between successive wavefronts. A moving observer, by contrast, changes the rate at which existing wavefronts are encountered.
For motion along the line connecting source and observer, define as positive when the observer moves toward the source and as positive when the source moves toward the observer. Both velocities are measured relative to the medium. In the ordinary subsonic case,
where is the observed frequency. An approaching source reduces the denominator, while an approaching observer increases the numerator. Either motion raises the measured frequency; recession lowers it. The formula assumes that the relevant velocities and wave speed remain effectively constant during the measurement.
The distinction between source and observer motion is physically important. A source moving toward a stationary observer at produces , or about . An observer moving toward a stationary source at the same speed measures . Equal closing speeds do not yield exactly equal shifts because source motion changes wavelength in the medium, whereas observer motion changes the rate of encountering that wavelength. A formula based only on relative speed is therefore generally insufficient for sound.
For example, take a -hertz source moving at metres per second toward a stationary observer in air with sound speed metres per second. The approaching frequency is approximately hertz. After the source recedes directly at the same speed, the frequency is approximately hertz. These values illustrate an idealized steady source, not the full sound of a real siren, whose emitted frequency and intensity may also vary.
The medium itself introduces additional structure. Sound speed depends on temperature and composition; in dry air near room temperature it is about metres per second. Wind transports the acoustic disturbance and changes propagation times. In a uniformly moving medium, the calculation can be performed in the medium’s rest frame, but spatially varying winds and temperature gradients bend acoustic paths through refraction. Water waves and waves in solids require their own propagation laws, especially where wave speed depends on frequency or direction.
Wavefront geometry, changing pitch, and supersonic motion
The elementary approach-and-recession description is a special case of a geometric problem. In three dimensions, the frequency shift depends on the components of source and observer velocity along the direction of wave propagation. For a nondispersive medium at rest, a useful instantaneous expression is
where points along the ray from the source’s emission event toward the observer’s reception event. The source velocity is evaluated at emission and the observer velocity at reception. This distinction matters because the source generally changes position while the wave is traveling.
An equivalent formulation follows from travel time. For a stationary observer and a point source in a uniform medium,
where is emission time, is reception time, and is the distance from the source at emission to the receiver. Differentiating gives
A decreasing emission distance compresses the arrival-time intervals, while an increasing distance stretches them. This formulation explains why an observer hears the source’s past motion rather than its position at the instant of reception.
A vehicle passing at a finite sideways distance does not, ideally, produce an instantaneous frequency jump. Its radial velocity changes continuously from approach to recession, so the observed pitch passes continuously through an intermediate value. The transition can nevertheless sound abrupt when the closest approach is small compared with the distances over which the vehicle is heard. Directional sound emission, reflections from buildings, and rapid changes in loudness can accentuate that impression. Closest approach at the time of emission must also be distinguished from closest approach at the later reception time.
As source speed approaches sound speed, wavefronts crowd together ahead of it. The subsonic formula’s denominator can approach zero, but this does not imply that a physical source produces an arbitrarily high, perfectly regular tone. Finite bandwidth, source size, nonlinear propagation, and the structure of the wavefront field become important. Above sound speed, successive wavefronts form an envelope called a Mach cone. Its half-angle satisfies
where is the Mach number. The associated sonic boom is a shock-wave phenomenon, not simply an unusually large pitch shift. It is generated along a supersonic trajectory, rather than only at the instant an aircraft first exceeds sound speed.
Relativistic Doppler shifts
Light in a vacuum does not propagate through a material medium whose rest frame determines its speed. Special relativity instead assigns the same vacuum light speed to every inertial reference frame. Source motion and observer motion cannot therefore be treated by simply replacing the sound speed with in the classical acoustic equation. The transformation of time intervals must also be included.
For a source and observer separating directly along their line of sight at relative speed , the observed frequency is
For direct approach, the ratio is inverted. Here is the frequency measured in the source’s rest frame. The expression combines the changing light-travel distance between successive emissions with the time dilation of the source clock. At low speeds, the longitudinal fractional shift is approximately for recession, but higher-order terms become significant at relativistic speeds.
An angular form makes the geometry explicit. Suppose the observer is at rest in a chosen inertial frame, the source moves with velocity , and the received light travels in direction , measured in that observer’s frame. Then
where . If the photon travels perpendicular to the source velocity in this frame, the result is : a transverse redshift. This is a characteristic relativistic effect, since the first-order classical contribution vanishes for that geometry.
“Transverse” must always be assigned to a specified frame. The apparent propagation direction of light changes between moving frames, a phenomenon called aberration. A ray perpendicular to the relative velocity in the observer’s frame is not perpendicular in the source’s frame. Consequently, apparently different transverse-Doppler formulas can describe different experimental geometries rather than contradictory predictions.
The same effect can be expressed using the energy of a photon, , where is the Planck constant. Observers moving relative to one another assign different energies and frequencies to the same photon. This is consistent with the transformation of energy and momentum under the Lorentz transformation. Relativistic frequency shifts are important in particle-beam experiments, high-speed astrophysical outflows, and precision measurements. In curved spacetime, the locally measured frequency additionally depends on the observer’s motion and the gravitational geometry along the signal’s path.
Astronomical spectroscopy and the interpretation of redshift
Astronomers commonly express a spectral displacement through the dimensionless redshift
where is the rest wavelength and is the observed wavelength. Positive denotes a shift toward longer wavelengths, while negative denotes a blueshift. For a small Doppler shift, a receding object’s radial velocity is approximately . If the displacement is entirely due to longitudinal special-relativistic motion,
so that
The approximation should not be used indiscriminately at large redshift.
The displacement of a spectral line measures motion along the line of sight, not the full three-dimensional velocity. Transverse motion usually requires an angular-motion measurement and a distance estimate. Spectroscopic velocities also depend on the chosen reference frame. Observations from Earth must account for the planet’s rotation and orbital motion, often reporting velocities relative to the solar system barycenter. At high precision, relativistic timing and gravitational corrections become relevant.
Periodic Doppler shifts reveal orbital motion in binary stars and the gravitational influence of planets on their host stars. The first widely recognized planet around a Sun-like star, 51 Pegasi b, was announced in 1995 by Michel Mayor and Didier Queloz using radial-velocity observations. For an approximately Keplerian orbit, the stellar velocity amplitude depends on the planet’s mass multiplied by the sine of the orbital inclination. Without additional information, spectroscopy generally gives a minimum planetary mass rather than a unique true mass.
Doppler measurements also reveal spatially distributed motion. The approaching and receding sides of a rotating galaxy shift spectral lines in opposite directions. Measurements of visible lines and the 21-centimetre line of neutral hydrogen trace galactic rotation. Across a rotating stellar disk, different surface regions contribute different velocities, broadening the integrated line profile even if the star’s center of mass has no changing radial velocity.
Not every astronomical redshift is a simple Doppler velocity. In cosmology, the expansion of the universe stretches wavelengths according to , where is the cosmic scale factor. General relativity also predicts gravitational frequency shifts between regions of different gravitational potential. Observed spectra may combine expansion, peculiar motion, gravitational effects, and shifts caused by atmospheric convection or line formation. Separating these contributions is essential when interpreting very small stellar velocities or very large cosmological redshifts.
Radar, weather observation, and spacecraft tracking
Radar transmits electromagnetic radiation and analyzes the returning echo. A moving target produces a Doppler shift on the outward and return paths. For a monostatic radar, in which transmitter and receiver are effectively at the same location, a target approaching radially at speed produces approximately
where is the transmitted wavelength. The factor of two distinguishes this reflection measurement from a simple one-way source-to-observer measurement. Recession reverses the sign of the shift.
A radar operating at a wavelength of centimetres receives a Doppler shift of about hertz from a target approaching at metres per second. Traffic radar, industrial speed sensors, and aircraft systems exploit this relation. An oblique measurement yields only the velocity component along the radar beam. A target moving at speed at angle to that beam gives a radial component . Bistatic systems, whose transmitter and receiver occupy different positions, require both propagation directions in their velocity relation.
Weather radar measures echoes from precipitation particles and other atmospheric scatterers. The phase change between successive pulses supplies the radial velocity of the scatterers. Spatial patterns of approaching and receding motion reveal wind shear, circulation, and storm organization. A single radar does not directly determine the complete wind vector at each point: its basic velocity observable is one component along the beam. Moreover, raindrops and ice particles fall relative to the surrounding air, so measured particle velocity is not always identical to air velocity.
Pulsed systems face a sampling limitation. If equally spaced pulses are transmitted at pulse-repetition frequency , the conventional unambiguous radial-velocity interval extends approximately to
A larger speed may produce an aliased value with an incorrect apparent magnitude or direction. Changing pulse intervals, combining measurements, and applying velocity-unfolding algorithms help address this problem, although practical systems must also manage range ambiguity and clutter.
Spacecraft tracking uses related principles at much greater distances. A stable ground frequency is transmitted, and a spacecraft transponder may return a signal with a prescribed frequency ratio. Comparing the received frequency with the expected reference measures changes in the radio path and constrains line-of-sight velocity. Precise interpretation accounts for station motion, spacecraft dynamics, light-travel time, plasma in the solar wind and ionosphere, and relativistic frequency changes. Doppler tracking complements signal-delay measurements, which constrain distance rather than directly measuring its rate of change.
Medical ultrasound and acoustic flow measurement
Doppler ultrasonography uses high-frequency sound to measure motion within the body, especially the movement of blood. An ultrasound transducer emits sound that is scattered by moving blood cells and tissues. Because the scatterers move relative to the transducer, the returning signal differs in frequency from the transmitted signal. At blood-flow speeds, which are very small compared with sound speed in tissue, the usual approximation is
where is the transmitted frequency, is scatterer speed, is the angle between motion and the beam, and is sound speed in tissue. Imaging systems commonly assume metres per second for soft tissue.
The factor of two again arises from the outward and return propagation paths. For a -megahertz beam, a scatterer moving at metres per second directly along the beam produces a shift of approximately kilohertz. The instrument detects this comparatively low-frequency difference by comparing the returning signal with an electronic reference. Thus, megahertz acoustic waves can encode velocity information in a signal within the audio-frequency range.
Continuous-wave Doppler transmits and receives continuously, allowing the detection of high velocities without the pulse-sampling ambiguity characteristic of pulsed instruments. Its limitation is that echoes arise from multiple locations along the beam, so depth is not uniquely selected. Pulsed-wave Doppler chooses a sample region using the echo’s arrival time, providing range localization. However, its measurable frequency shifts are constrained by the pulse-repetition frequency and can alias when that sampling limit is exceeded.
Spectral Doppler displays the distribution of frequency shifts over time, representing a range of velocities within the sampled region. Color Doppler maps estimated signed velocity information onto an image, while power Doppler displays a measure related to the strength of the Doppler signal rather than its signed mean velocity. Display colors are conventions established by the instrument’s settings; they do not inherently identify arteries and veins or oxygenated and deoxygenated blood.
The principal geometric limitation is angle dependence. Motion perpendicular to the beam produces no first-order Doppler shift, and a small angular error can produce a large inferred-speed error when the beam is nearly perpendicular to flow. Finite sample size, vessel motion, scattering characteristics, and instrument settings also affect the recorded spectrum. Similar acoustic principles underlie underwater current profilers, which infer water motion from echoes scattered by suspended particles. In both medical and environmental applications, the measured quantity is initially the motion of acoustic scatterers along a propagation direction.
Atomic motion, spectral broadening, and laser techniques
A Doppler measurement need not show a single displaced spectral line. If an ensemble contains emitters or absorbers with different velocities, their individual shifts combine into Doppler broadening. In a dilute gas at thermal equilibrium, the velocity component along any chosen direction follows a Gaussian distribution. At nonrelativistic temperatures, a narrow atomic transition therefore acquires an approximately Gaussian Doppler profile centered near its rest frequency.
For atoms of mass , temperature , and transition frequency , the Doppler full width at half maximum is
where is the Boltzmann constant. The width grows as , decreases as , and is proportional to the transition frequency. Thermal motion can consequently obscure fine spectral structure even when the source and apparatus have no relative bulk motion.
Real spectral lines also have natural broadening associated with finite excited-state lifetimes and collision broadening associated with interactions among particles. These mechanisms are not Doppler effects. A commonly used model combines a Gaussian Doppler contribution with a Lorentzian contribution to form a Voigt profile. Distinguishing broadening mechanisms makes it possible to infer temperatures, densities, and velocity distributions from laboratory and astronomical spectra, although unresolved flows or turbulence can imitate part of the thermal width.
Lasers allow experiments to select or manipulate particular velocity classes. In saturated-absorption spectroscopy, counterpropagating beams can interact with the same atoms under conditions that reveal narrow features within a broad thermal profile. Such methods provide frequency references and resolve structures hidden by ordinary Doppler broadening. Arrangements using counterpropagating photons can also cancel first-order Doppler contributions in particular two-photon transitions. The extent of cancellation depends on the actual beam geometry and frequencies.
The Doppler effect is also central to laser cooling. An atom moving toward a laser beam sees its light shifted to a higher frequency. If opposing beams are tuned slightly below an atomic resonance, an atom moving toward one beam can preferentially absorb photons from it and receive momentum opposing its motion. Repeated absorption and spontaneous emission produce a velocity-dependent damping force. Random photon recoils simultaneously cause heating; their balance sets the standard Doppler cooling limit for an idealized two-level transition. More elaborate cooling mechanisms can reach lower temperatures.
In engineering, laser Doppler velocimetry measures particle or surface motion from optical frequency differences. Crossing laser beams can create an interference fringe pattern; particles crossing those fringes scatter light with an intensity modulation related to their velocity. This supports noncontact measurements in fluid mechanics and vibration studies.
Signal analysis, precision, and limits of interpretation
A Doppler shift is fundamentally a change in the rate at which wave phase accumulates. If a received signal has phase , its instantaneous frequency is
Comparing the received signal with a reference converts a small frequency difference into a slowly varying beat signal. This principle underlies coherent radar, optical heterodyne detection, radio tracking, and many ultrasound systems. In practice, instruments often measure phase differences first and infer frequency or velocity from their change over time.
Finite observation time limits frequency discrimination. For a simple Fourier analysis over a duration , the characteristic spectral spacing is of order . Longer integration can distinguish smaller frequency differences, but only if the signal remains coherent and sufficiently stable. Acceleration, vibration, or changing propagation conditions can turn a narrow spectral peak into a broadened or time-dependent feature. More sophisticated estimation methods can outperform a naive spectral-bin estimate under specified signal and noise models, but they do not remove uncertainty from finite, noisy data.
The transmitter’s frequency stability and the receiver’s reference are as important as spectral resolution. Oscillator drift can imitate target motion, while optical-spectrograph changes can imitate a stellar velocity shift. Precision work therefore depends on calibration, stable references, and a defensible account of measurement uncertainty. In high-precision stellar spectroscopy, atmospheric absorption, illumination changes, and stellar surface activity must also be distinguished from orbital motion.
A single frequency shift rarely determines a complete physical state. It generally constrains a projection of velocity, and multiple moving components may overlap in one spectrum. Reflected signals depend on both illumination and observation geometry. In dispersive media, phase velocity and group velocity differ, so carrier-frequency shifts and pulse-arrival timing require separate treatment. A time-varying medium can itself alter a signal’s frequency, even without a simple moving point source or receiver.
Some structured waves support a rotational Doppler effect, in which rotation couples to angular phase structure rather than only to linear motion along a ray. Such measurements can probe rotation using optical modes carrying angular momentum. They extend the phase-based description but require a specified wave mode and geometry.
The relation between Doppler shift and amplitude also needs care. A passing source becomes louder and then quieter partly because its distance and directional emission change, not merely because its frequency shifts. For relativistic radiation, frequency changes accompany transformations of arrival rate and angular distribution, producing beaming effects. Frequency, intensity, travel time, and apparent direction are related observables, but each carries distinct information and requires its own propagation model.