Cross-correlation function
A weighted line mask compresses thousands of spectral lines into an average profile. It is fast and interpretable, but the mask, rejected regions and profile fit define the measurement.
Baranne et al. (1996)Independent Literature Synthesis · Precision Spectroscopy
Precision RV Spectrograph Literature Atlas
A paper-by-paper account of optical and near-infrared Doppler instruments: how each system is configured, what its authors tested, what the measurements establish, what remains difficult, and which next checks may be useful.
A careful boundary. A design goal, calibration test, formal uncertainty and stellar time-series RMS are different measurements. “Research note” means I reviewed the listed sources; it does not mean every paper ever written about that instrument.
Landscape at a glance
Where precision-RV spectrographs sit
This two-panel atlas joins nominal wavelength grasp to the velocity scale reported in each source, while keeping on-sky measurements separate from requirements, commissioning and subsystem tests.
Begin here
From starlight to a velocity time series
A five-step measurement chain, the equations behind it, and the places where apparently similar spectrographs become scientifically different.
Interactive worked example
Make the invisible shift visible
Move a star toward or away from us and see how a radial velocity becomes a wavelength displacement. The track is drawn in velocity space so tiny shifts stay visible.
Model boundary. This uses the first-order relation Δλ/λ ≈ v/c and assumes effectively infinite signal-to-noise and perfect detector sampling. It illustrates scale only; it does not model line shape, photon noise, stellar variability, calibration error or achieved instrument precision.
Negative values produce a blueshift; positive values produce a redshift.
Try a scale
Line position in radial-velocity space
Observed line is 1.0 m/s to the red of the rest line.
- Shifted wavelength
- 550.000001835 nm
- Wavelength shift
- +0.001835 pm
- Fractional shift
- +3.336 × 10⁻⁹
- At R = 100,000
- 0.0334% of one resolution element
The calculation follows the same first-order Doppler relation introduced above. For the controlled line-width and resolving-power experiment, see the Doppler-information study.
Calibration and extraction
Three common measurement architectures
From extracted spectrum to velocity
The pipeline is part of the measurement
A stable optical bench is necessary, but the reported radial velocity also depends on how pixels, wavelengths, tellurics, templates and timestamps are modelled.
Template matching and forward models
Least-squares methods compare each observation with a high-signal empirical template; broader forward models can also fit the wavelength solution, line-spread function, calibration and contamination. More retained information also means more assumptions to validate.
Zechmeister et al. (2018)Reference frame and exposure time
Earth's orbital and rotational motion is removed with a relativistic barycentric correction. At EPRV scales, the correction must follow the photon-weighted exposure—often as a function of wavelength—not just the shutter midpoint.
Wright & Eastman (2014) · Tronsgaard et al. (2019)Detector and atmosphere
Pixel stitching, charge transfer, persistence, non-linearity and micro-tellurics can move or reshape lines. Calibration cannot automatically remove effects that do not share the stellar light path or the same detector response.
Dumusque et al. (2015)Read the extraction and validation note with primary sources
Scope boundary
Not every spectrograph is an EPRV instrument
| Class | Examples | Measurement emphasis | In directory? |
|---|
Location guide
Where the instruments collect light
Each point uses a cited site or telescope coordinate. Several instruments at one observatory share a marker; this is a reading map, not a survey-grade geodetic product.
90° N
90° S
180° W
180° E
Ground facility◆Space spectrographs appear as comparator notes, not ground-map points
Base geography: Natural Earth 1:110 million, public-domain vector data rendered with a UN-style boundary view. Boundaries provide geographic context and do not imply a position on disputed territory.
Browse the full facility index
Instrument Directory
One Record per Physical Spectrograph
The directory includes dedicated precision-RV systems, general high-resolution instruments with RV measurements, historical prototypes, and funded or commissioning projects.
Evidence is field-specific: a paper can confirm that an instrument exists without proving its present operating state or exact location. Open fields remain visible in each note.
Numbers from Published Studies
Compare Specifications Only Within Context
The performance layer contains 21 instruments and 22 source-linked numerical records because EXPRES has separate calibration and on-sky rows. Wavelength limits and resolving power are representative modes, not acceptance-test values.
| Instrument | Facility | Coverage | R | Status | Evidence class |
|---|
Read the number in context
What did each number measure?
Open a card to see its target, baseline, limitation and primary paper.
Goal or requirementCalibrationOn-sky result
Paper-based interpretation
Nine coupled error domains
Each section states the mechanism, a published example, common mitigations, the residual problem, and links to the papers.
Original Figures
Context Before Comparison
Figures are generated from the committed tables. Select any figure to inspect the full-resolution version; no third-party paper figure is reproduced.
How This Was Made
Every Number Points Back to a Source
Each numerical row records the metric, measurement context, target or sample, baseline, limitation, source URL and access date. Automated checks reject unknown instruments, duplicate identifiers and rows without a stated limitation.