Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Virgo now unbias its reconstructed strain online, cutting residual bias below 2% in modulus and 15 mrad in phase across 10 Hz–1 kHz during O4.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-03 21:58 UTC pith:BPQRJ4FO

load-bearing objection Frequency-dependent calibration bias/uncertainty is a real O4 improvement, but the full-band precision claim rests on self-referential validation and an unverified 10–18 Hz extrapolation. the 4 major comments →

arxiv 2511.12566 v1 pith:BPQRJ4FO submitted 2025-11-16 physics.ins-det astro-ph.IM

From the Virgo interferometer calibration to the bias and uncertainty of the h(t) detector strain during the O4 run

Cervane Grimaud , Florian Aubin , Beno\^it Mours , Thierry Pradier , Lo\"ic Rolland , Monica Seglar-Arroyo , Hans Van Haevermaet , Pierre Van Hove
show 1 more author
Didier Verkindt
This is my paper
classification physics.ins-det astro-ph.IM PACS 04.80.Nn95.55.Ym
keywords gravitational waveVirgocalibrationh(t) strainbiasuncertaintyO4 runphoton calibrator
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper describes how the Virgo gravitational-wave detector calibrates its measurement chain and, for the first time in an observing run, actively corrects the reconstructed strain signal h(t) online rather than just estimating a single, frequency-independent uncertainty. The central achievement is a new method that uses two sets of injected calibration lines—weekly lines for bias, permanent lines for uncertainty—to build a frequency-dependent bias and uncertainty curve. Applying this bias correction reduces the residual reconstruction error to below 2% in modulus and below 15 mrad in phase over most of the 10 Hz–1 kHz band. The result matters because accurate, low-bias strain reconstruction is essential for extracting weak gravitational-wave signals and for correctly estimating source parameters.

Core claim

The paper claims that a newly developed frequency-dependent bias and uncertainty computation method, applied online, successfully unbias the reconstructed h(t) strain during Virgo's O4 run. Using weekly injected calibration lines, the method estimates the reconstruction bias and corrects h(t) in real time, resulting in a residual bias below 2% in modulus and below 15 mrad in phase over 10 Hz–1 kHz. Using permanent lines, the same method computes a frequency-dependent reconstruction uncertainty, replacing O3's frequency-independent uncertainty. The overall strain reconstruction uncertainty is reported below 2.5% in modulus in the 10 Hz–1 kHz band and within 25 mrad in phase below 200 Hz, risi

What carries the argument

The key machinery is a Monte-Carlo-style interpolation scheme: from the measured modulus and phase deviations at injected calibration lines (27 weekly lines for bias, 16 permanent lines for uncertainty), the method draws random samples from Gaussian distributions of the line data, linearly interpolates across the 10 Hz–10 kHz band, and uses the resulting distributions to compute a per-frequency-bin mean (bias) and standard deviation (uncertainty). This frequency-dependent curve is then applied online to unbias h(t).

Load-bearing premise

The extrapolation of the bias and uncertainty below the first injected calibration line (down to 10 Hz) assumes that the frequency-dependent bias shape does not change abruptly below 18 Hz, which is uncertain due to possible low-frequency control-loop or suspension effects.

What would settle it

Inject a dense set of calibration lines in the 10–30 Hz band (or perform a broadband injection) and compare the resulting measured bias and uncertainty to the values extrapolated by the current method; if the difference exceeds the reported 2% modulus or 15 mrad phase, the extrapolation assumption is disproven.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • If the claimed accuracy holds, gravitational-wave events detected by Virgo during O4 will have strain calibration errors below the stated levels, improving the reliability of source parameter estimation.
  • The online unbiasing reduces a systematic that previously had to be accounted for in post-processing, simplifying downstream analysis and potentially reducing systematic biases in astrophysical inferences.
  • The method's replacement of a frequency-independent uncertainty with a frequency-dependent one allows better exploitation of the sensitive band, especially at low frequencies where the old uncertainty was overly conservative.
  • Monthly updates of the bias and uncertainty enable monitoring of long-term drift, with the paper reporting the bias was stable, changing only twice over 16 months.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same interpolation-based method could be applied to other interferometers or future detectors (e.g., Einstein Telescope or Cosmic Explorer) to provide robust, data-driven uncertainty estimation without relying on dense line injections.
  • The claim of stability over 16 months suggests that the bias is dominated by slow, well-modeled effects, implying that frequent recalibration may be unnecessary; a longer trending study could identify when bias changes occur and what causes them.
  • The method's reliance on a small number of injected lines implies that its accuracy in regions with complex transfer-function features (e.g., suspension resonances or control loop notches) depends on line placement; testing with denser injections in those regions would probe the interpolation's validity.
  • If the extrapolation below the lowest line down to 10 Hz is validated by independent measurements (e.g., NCal injections or broadband injections), the method could be extended to the full LIGO-Virgo-KAGRA low-frequency band, which is important for binary neutron star and intermediate-mass black hole searches.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. This paper describes the calibration chain for the Virgo interferometer during the O4 run, with emphasis on the Photon Calibrator (PCal) as a displacement reference, the calibration of the electromagnetic actuators, and the reconstruction of the detector strain h(t). The main new element is a frequency-dependent bias and uncertainty estimation method that uses permanent and weekly calibration lines, with linear interpolation between line frequencies and linear extrapolation below the lowest line, applied online to unbias h(t). The authors report a residual bias below 2% in modulus and 15 mrad in phase over 10 Hz–1 kHz, a frequency-dependent uncertainty below 2.5% in modulus, and claim in the abstract an overall precision of 2% in modulus and 30 mrad in phase on the 10 Hz–2 kHz band.

Significance. If the method is valid, it represents a practical operational improvement over the O3 frequency-independent calibration, enabling online unbiasing of h(t) and a frequency-dependent uncertainty estimate. The paper draws on real O4 data and describes a computationally simple Monte Carlo interpolation scheme that could be useful to the LIGO-Virgo-KAGRA calibration community. However, the evidence presented is not yet sufficient to support the strongest claims: the verification relies on the same injection lines used to construct the bias, and the low-frequency extrapolation is an unvalidated assumption.

major comments (4)
  1. [§3 and Fig. 1] The residual-bias verification in Fig. 1 is self-referential. The h_unbias/h_inj measurement is computed from the same weekly injection lines that are used to build the bias estimate in §3. A value near unity at these frequencies confirms only that the interpolation reproduces the average of the input points; it does not test the bias at inter-line frequencies or below 18 Hz. The claim that 'the unbiasing allows to get a reconstructed strain with a residual bias within less than 2% over 10 Hz to 1 kHz' is therefore stronger than the evidence shown. Please restrict the residual-bias statement to the line frequencies or validate it with an independent reference, such as the broadband h_rec/h_inj injections mentioned in §4.
  2. [§3, low-frequency extrapolation] The extrapolation from the lowest two weekly lines (the lowest being 18 Hz) down to 10 Hz assumes that the bias trend continues linearly. Because the first line lies at 18 Hz and the observation band extends to 10 Hz, the 10–18 Hz sub-band is outside the interpolation range. Suspension resonances or control-loop features below 18 Hz could produce curvature that the first-order extrapolation misses. The permanent lines include a 16 Hz line; using it as a cross-check of the extrapolated bias would provide a direct test. As written, the claimed 2%/15 mrad residual bias in the 10–18 Hz sub-band is unsupported.
  3. [Abstract vs §4 and Conclusion] There is an inconsistency between the abstract's claim of '2% in modulus and 30 mrad in phase' over the '10 Hz to 2 kHz band' and the results reported later. Section 4 gives residual bias below 2% over 10 Hz–1 kHz and below 1% over 30–500 Hz, while the conclusion states an overall uncertainty below 2.5% in modulus in the 10 Hz–1 kHz band, with phase values only below 200 Hz. No results are shown above 1 kHz. The high-frequency behavior in §3 is set to modulus 1 and phase 0 at 10 kHz, which is an arbitrary boundary condition, not a measurement. Please clarify which quantity is claimed for which frequency band, and provide supporting data for the 1–2 kHz portion of the abstract's claim.
  4. [§4, uncertainty combination] The construction of the total uncertainty is not sufficiently specified to be reproducible. The PCal uncertainty appears as 0.48% in §1 but as a 'conservative' 0.6% in §4; the actuator-model uncertainties (0.5% modulus, 5 mrad phase) are quoted without derivation; and the 'additional errors around 50 Hz and 150 Hz' are introduced after broadband injections without any quantitative description of how their amplitudes were obtained or how they are combined with the other terms. Without this information, the final uncertainty curve (Fig. 2) and the 'below 2.5%' statement cannot be independently checked.
minor comments (5)
  1. [Eq. (1)] Equation (1) is typeset in a garbled way, with 'l vp' and 'l M' subscripts and the placement of S^{-1}_{PCal}(f) unclear. Please define all symbols explicitly and present the equation in standard form.
  2. [§2] The quantity h_inj is used in the monitoring ratio h_raw/h_inj but is never explicitly defined. State how h_inj is computed from the actuator model and why it is a convenient reference.
  3. [§2–§3] The permanent lines are described as 'a set of 16 lines' but their frequencies are not listed. Since the uncertainty method in §3 relies on these lines, a table or list of their frequencies would help the reader follow the method.
  4. [Fig. 1] Figure 1 is referenced neither in the §4 text nor elsewhere. The caption labels panels (a) and (b), but the text does not call them out; please add an explicit reference and describe the curves.
  5. [Abstract] The abstract contains a sentence fragment: 'Before explaining how the Photon Calibrator is used to calibrate every Virgo mirror actuators.' This should be rephrased.

Circularity Check

1 steps flagged

Residual-bias verification reuses the same weekly lines used to build the frequency-dependent bias; the reported <2% residual is partly self-consistency, though the absolute calibration chain uses independent NCal/PCal references.

specific steps
  1. fitted input called prediction [Section 3 (method) and Section 4 (results), Fig. 1 and caption]
    "Using the weekly injections we are able to compute a bias estimation which is then used to correct online the reconstructed strain h raw to get the unbiased one h unbias. We can monitor the unbiasing by doing an h unbias/hinj measurement which will show the residual bias of the reconstruction. ... In figure 1 we see that the unbiasing allows to get a reconstructed strain with a residual bias within less than 2% over 10 Hz to 1 kHz and less than 1% over 30 to 500 Hz for the modulus."

    The bias estimate is built from the same weekly-line h_raw/h_inj data that are then used for verification: Section 3 averages those line values and linearly interpolates between the 27 points. Since h_unbias = h_raw / bias_est, the displayed statistic is h_unbias/h_inj = (h_raw/h_inj)/bias_est. At the weekly-line frequencies bias_est is fit to h_raw/h_inj by construction, so a residual near 1 is expected and not an independent confirmation. Fig. 1 explicitly compares h_raw/hinj and h_unbias/hinj 'performed with the same average period.' Thus the claimed 2% precision is partly a self-consistency check at the fitted points, while inter-line and 10-18 Hz extrapolated behavior is not independently validated.

full rationale

The calibration chain itself is not circular: PCal displacement comes from a measured power and an independent mechanical model, EM actuator responses are obtained by transfer-function ratios against that PCal reference, and the final uncertainty adds independent PCal and actuator-model uncertainties in quadrature. The references to prior Virgo papers are normal calibration documentation, not a load-bearing uniqueness theorem or an ansatz smuggled in via citation. The main circular element is the demonstration of the new unbiasing: the same weekly lines that define the bias are reused to display h_unbias/h_inj residuals, so the quoted precision is partly forced at the fitted line frequencies. The paper itself concedes a limitation at 50/150 Hz ('additional errors ... were added after looking at broadband injections'), showing that line-based interpolation can miss real features, yet no analogous full-band broadband check is shown for inter-line and 10-18 Hz frequencies. Because the central method is a calibration cross-check with independent absolute references rather than a parameter-free prediction derived from fitted inputs, the circularity is partial rather than total.

Axiom & Free-Parameter Ledger

5 free parameters · 5 axioms · 0 invented entities

The central claim rests on a set of analysis choices (monthly averaging, N=1000, 0.125 Hz binning, low-frequency extrapolation) and on the validity of the calibration reference models. No new physical entities are introduced. The method is essentially a statistical calibration technique, so the burden on free parameters and assumptions is moderate.

free parameters (5)
  • Averaging period = one month
    Chosen by hand; the bias and uncertainty are averaged over monthly chunks. The sensitivity to this choice is not assessed.
  • Number of random draws (N) = 1000
    Used for sampling from the Gaussian distributions to build per-frequency distributions. Not varied or justified.
  • Frequency bin width = 0.125 Hz
    Linear interpolation is evaluated on a 0.125 Hz grid; this step size is a free choice.
  • Low-frequency extrapolation slope = from first two lines
    The bias below the first weekly line (18 Hz) is extrapolated using the slope of the two closest lines; this is an ad hoc assumption.
  • Additional 50/150 Hz errors = values not given
    Errors at the power-line frequency and its second harmonic were added after inspecting broadband injections; the magnitude is not specified in this paper.
axioms (5)
  • domain assumption The injected calibration lines provide an unbiased reference of the true applied strain (h_inj is accurate).
    The bias computation assumes that h_inj, computed from the actuator response model, faithfully represents the physical displacement; any error in the actuator model is absorbed into the measured bias and propagated to the unblased strain.
  • domain assumption The scatter of line measurements follows a Gaussian distribution.
    The method constructs Gaussian distributions from mean and error values and samples from them; this is not validated in the paper.
  • ad hoc to paper Linear interpolation between injected lines is a sufficient representation of the frequency-dependent bias and uncertainty.
    The method linearly interpolates between lines, which assumes the bias varies smoothly; this may miss narrow features between lines.
  • domain assumption The calibration uncertainties (PCal 0.6%, actuator 0.5%, 5 mrad) are correct and add in quadrature.
    These values are cited to internal theses/VIR documents and are not re-derived here; they are treated as known inputs.
  • domain assumption The detector's calibration state is stable over the monthly averaging period.
    The method assumes that averaging over a month gives a meaningful estimate of bias and uncertainty; if the state drifts within the month, the result is a mixture.

pith-pipeline@v1.3.0-alltime-deepseek · 4460 in / 8961 out tokens · 76807 ms · 2026-08-03T21:58:54.932273+00:00 · methodology

0 comments
read the original abstract

Since the first gravitational wave detection in 2015, ground-based interferometer sensitivities have significantly improved, requiring highly precise calibration to ensure accurate reconstruction of the h(t) strain signal. In this talk we will outline the Virgo interferometer calibration steps performed in preparation of the O4b run started in April 2024. We will first describe the Photon Calibrator power devices intercalibration allowing for a 0.48% precision on mirror displacement. Before explaining how the Photon Calibrator is used to calibrate every Virgo mirror actuators. We will also discuss the monitoring of the h(t) strain reconstruction during the run showing that, on the 10 Hz to 2 kHz band, the reconstructed strain achieves a precision of 2% in modulus and 30 mrad in phase. Special emphasis will be given on the newly developed frequency-dependent bias and uncertainty computation method and the resulting online unbiasing of the h(t) strain.

Figures

Figures reproduced from arXiv: 2511.12566 by Beno\^it Mours, Cervane Grimaud, Didier Verkindt, Florian Aubin, Hans Van Haevermaet, Lo\"ic Rolland, Monica Seglar-Arroyo, Pierre Van Hove, Thierry Pradier.

Figure 1
Figure 1. Figure 1: Comparison between the hraw/hinj measurements performed using an average of the weekly lines data from the 28th of January to the 25th of February 2025 and the hunbias/hinj measurements performed with the same average period. The goal for O4 was to be able to compute a frequency-dependent reconstruction uncertainty. This is performed using the exact same method as for the bias computation but using the unb… view at source ↗
Figure 2
Figure 2. Figure 2: Result of the reconstruction residual bias [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. GW240925 and GW250207: Astrophysical Calibration of Gravitational-wave Detectors

    gr-qc 2026-05 unverdicted novelty 8.0

    The first informative astrophysical calibration of gravitational-wave detectors is reported using GW240925 and GW250207.

Reference graph

Works this paper leans on

12 extracted references · cited by 1 Pith paper

  1. [1]

    The foundation of the general theory of relativity

    Albert Einstein. The foundation of the general theory of relativity. Annalen Phys., 49(7):769–822, 1916

    1916

  2. [2]

    B. P. Abbott and all. Prospects for observing and localizing gravitational-wave transients with Advanced LIGO, Advanced Virgo and KAGRA. Living Reviews in Relativity, 23(1), September 2020

    2020

  3. [3]

    Advanced Virgo: a second-generation interferometric gravitational wave detector

    F Acernese and all. Advanced Virgo: a second-generation interferometric gravitational wave detector. Classical and Quantum Gravity, 32(2):024001, dec 2014

    2014

  4. [4]

    Advanced Virgo plus phase 1, Design report, 2019

    Virgo collaboration. Advanced Virgo plus phase 1, Design report, 2019. VIR-0596A-19, https://tds.virgo-gw.eu/?r=15777

    2019

  5. [5]

    Newtonian calibrator tests during the Virgo O3 data taking

    D Estevez, B Mours, and T Pradier. Newtonian calibrator tests during the Virgo O3 data taking. Classical and Quantum Gravity, 38(7):075012, March 2021

    2021

  6. [6]

    The Virgo Newtonian calibration system for the O4 observing run

    F Aubin, E Dangelser, D Estevez, A Masserot, B Mours, T Pradier, A Syx, and P Van Hove. The Virgo Newtonian calibration system for the O4 observing run. Classical and Quantum Gravity, 41(23):235003, October 2024

    2024

  7. [7]

    Estevez, P

    D. Estevez, P. Lagabbe, A. Masserot, L. Rolland, M. Seglar-Arroyo, and D. Verkindt. The Advanced Virgo Photon Calibrators. Classical and Quantum Gravity, 38:075007, 2021

    2021

  8. [8]

    Improving the Advanced Virgo+ calibration with the Photon Calibrator for the 04 run

    Paul Lagabbe. Improving the Advanced Virgo+ calibration with the Photon Calibrator for the 04 run. PhD thesis, Universit´ e Savoie Mont Blanc, 2023

    2023

  9. [9]

    Virgo interferometer calibration and h(t) detector strain reconstruction uncertainty computation during the O4 observing run

    Cervane Grimaud. Virgo interferometer calibration and h(t) detector strain reconstruction uncertainty computation during the O4 observing run. PhD thesis, Universit´ e Savoie Mont Blanc, 2025

    2025

  10. [10]

    Virgo PCal intercalibration for the O4b run, 2024

    Cervane Grimaud and Loic Rolland. Virgo PCal intercalibration for the O4b run, 2024. VIR-0856A- 24

    2024

  11. [11]

    Virgo PCal power calibration for O4b run, 2024

    Cervane Grimaud and Lo ¨ ıc Rolland. Virgo PCal power calibration for O4b run, 2024. VIR-0856A-24

    2024

  12. [12]

    Calibration of Advanced Virgo and reconstruction of the detector strain h(t) during the observing run O3

    F Acernese and all. Calibration of Advanced Virgo and reconstruction of the detector strain h(t) during the observing run O3. Classical and Quantum Gravity, 39(4):045006, January 2022

    2022