{"id":"1f381b68-19ec-4b19-8129-ba204cc8c408","arxiv_id":"2411.14191","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Tilt-to-length coupling noise can be fitted and subtracted from simulated LISA data without being confused by, or damaging, gravitational wave signals from several source types.","lead":"Noise from tiny spacecraft wobbles can leak into LISA's laser measurements. This paper shows, using simulated data, that removing that noise still works when real gravitational wave signals are present.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'GW signal characteristics were not altered' claim is supported only by visual ASD overlap in two one-day cases; no parameter-estimation or phase-coherent test is reported, and the paper explicitly defers this to future work.","rationale":"Good-faith reading: the paper does what it says in the body: it simulates one-day LISA data with five GW configurations (verification binaries, SGWB, galactic binaries, MBHB, full Sangria) and shows that MCMC TTL coefficient estimates stay below 0.1 mm/rad and residuals after subtraction meet the LISA requirement. The controlled comparison with and without GW signals using identical noise realizations is a sensible design, and the use of established tools (LISANode, PyTDI, LDC Sangria, LISA GW Response) is a real strength. My concern is not with the coefficient-estimation claim within the stated simulation setup; it is with the abstract's unqualified assertion that GW signal characteristics are not altered. The body only demonstrates spectral-level residual agreement in two one-day cases and explicitly defers the question of longer integration and parameter-estimation effects to future work. Because the stated goal of distinguishing TTL from GW is ultimately about protecting astrophysical inference, the missing quantitative signal-recovery test is load-bearing: a bias in TTL coefficients that is too small to appear in an ASD could still shift parameter posteriors. The white-jitter limitation is acknowledged by the authors and is a second reason the in-flight claim is not yet established. Both concerns point to the same remedy: rerun with realistic colored jitter and perform end-to-end parameter estimation. I agree with the reader's CONDITIONAL verdict; no change is needed.","tokens_in":14148,"tokens_out":7726,"duration_ms":79524,"concrete_test":"Re-run the Section V.D MBHB2 scenario through the full pipeline, then feed the TTL-subtracted TDI AET channels to a standard LISA parameter-estimation code (e.g., the LDC global fit or bilby) and compare recovered chirp mass, mass ratio, coalescence time, luminosity distance, and sky location with the injected Sangria values; quantify any bias in units of posterior width. Separately, replace the white SC/MOSA jitter with a colored jitter realization from a LISA dynamics/control-loop simulation (e.g., [38,39]) and repeat the coefficient estimation. If recovered parameters remain unbiased and coefficient errors stay below 0.1 mm/rad under colored jitter, the concern is resolved; otherwise the abstract's 'not altered' claim must be qualified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section V.C (fig. 9) and V.D (fig. 14) are the only places where the paper tests whether TTL subtraction preserves GW signals. The test is a visual overlay of the ASD of the injected GW response with the residual obtained by subtracting the fitted TTL model from data that contain no instrument noise other than TTL. An ASD match cannot certify that signal characteristics are unaltered: phase-coherent distortions, time shifts, or spectral leakage can leave an ASD nearly unchanged while biasing astrophysical parameter estimates. The paper's own Summary states: 'the latter was shown only indirectly in figures 9 and 14 for one day of data. The results would need to be confirmed considering the longer integration times for GW analyses and their ultimate effect on the astrophysical parameter estimation.' The abstract, however, states without qualification that 'the GW signal characteristics were not altered by the TTL coupling subtraction.' A second, related gap is the white-jitter assumption: Appendix A uses jitter that is white in the 3 mHz–0.9 Hz fit band, and Section V.D concedes 'We cannot say, whether or how other jitter shapes would affect the findings of this paper.' Since real LISA jitter is expected to be colored, the fit behavior demonstrated here is not yet the behavior that will be needed in flight.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates whether post-processing subtraction of tilt-to-length (TTL) noise in LISA remains accurate and non-destructive when gravitational wave (GW) signals are present in the data. The authors simulate one day of LISA data with LISANode, include four classes of GW signals (verification binaries, a stochastic GW background, galactic white-dwarf binaries, massive black hole binary mergers, and a full LDC Sangria combination), and use an MCMC fit to estimate the 24 TTL coupling coefficients in the TDI AET variables. They report that coefficient estimation errors remain below the 0.1 mm/rad requirement, that residual noise after subtraction stays below the LISA requirement, and that visual overlays of amplitude spectral densities suggest the GW signal is not degraded by the subtraction. The manuscript also confirms that static-arm simulations are adequate despite small real arm-length variations.","tokens_in":14417,"tokens_out":3741,"duration_ms":36531,"significance":"If the central claim holds, the paper provides valuable evidence for the LISA data-processing pipeline: TTL noise subtraction should not be catastrophically confused with GW signals, and the subtraction itself should not corrupt the GW response. The study's strengths include the use of realistic LISANode simulations, public LDC data, multiple GW source types, and an MCMC estimator that is not given the true coupling coefficients, so the coefficient-error results are a genuine test of the estimation procedure. The residual ASDs and coefficient-error plots are informative and support the main quantitative claims about noise subtraction. However, the claim that 'GW signal characteristics were not altered' is supported only by qualitative ASD overlays over one day of data, which is not sufficient for the strength of the abstract statement.","major_comments":[{"comment":"The abstract's claim that 'the GW signal characteristics were not altered by the TTL coupling subtraction' is not supported by the evidence presented. In Fig. 9 and Fig. 14 the test is a visual overlap between the ASD of the injected GW response (purple) and the residual obtained by subtracting the fitted TTL model from data with no instrument noise other than TTL (yellow). An ASD match is insensitive to phase-coherent distortions, time shifts, or spectral leakage, any of which could bias astrophysical parameter recovery while leaving the ASD nearly unchanged. The Summary itself concedes that the signal-preservation point 'was shown only indirectly in figures 9 and 14 for one day of data' and that confirmation with longer integration times and parameter estimation is required. Please either add a phase-coherent or parameter-estimation test on the residual data, or qualify the abstract and conclusions accordingly.","section":"Section V.C, V.D and Summary"},{"comment":"The simulations assume jitter that is white in the 3 mHz to 0.9 Hz fit band (Appendix A, Eqs. (A1)-(A3)), while real LISA jitter is expected to be colored. Section V.D states: 'We cannot say, whether or how other jitter shapes would affect the findings of this paper.' Because the central claim concerns in-flight applicability, this is a substantial limitation. The abstract should not present the results without this caveat. Either demonstrate robustness to a colored-jitter model (e.g., by repeating the coefficient-fit and subtraction tests with a roll-off or otherwise colored spectrum) or explicitly scope the headline claim to the white-jitter case.","section":"Abstract and Section V.D"},{"comment":"All signal-preservation tests are performed on one day of data. The claim that GW signals are not altered is therefore limited to a one-day timescale; the Summary acknowledges that longer integration times and the ultimate effect on astrophysical parameter estimation remain open. Since LISA science analyses will integrate for weeks to months, the manuscript should either extend the analysis to a longer data stretch or at least move the one-day limitation into the abstract so that the strength of the claim matches the evidence.","section":"Section V.C, V.D and Summary"}],"minor_comments":[{"comment":"The heading 'TIL T-TO-LENGTH NOISE IN LISA' contains a typo; it should read 'TILT-TO-LENGTH NOISE IN LISA'.","section":"Section II heading"},{"comment":"In the sentence introducing the interferometer noises, 'lenght readout' should be 'length readout'.","section":"Appendix A"},{"comment":"The caption spells 'Keplarian' orbits; the standard spelling is 'Keplerian'.","section":"Figure 2 caption"},{"comment":"The verification binary 'HMCnc' is presumably 'HM Cnc' (the cataclysmic variable HM Cancri); please use the standard astronomical designation.","section":"Section V.A"},{"comment":"The phrase 'The here presented results cannot directly be compared with [10]' would read more naturally as 'The results presented here cannot be directly compared with [10]'.","section":"Section V.D"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid, well-scoped simulation study that will be of interest to the LISA data-analysis community. The main issue is the mismatch between the unqualified abstract claim and the actual evidence, which is fixable either by adding a quantitative signal-preservation test (e.g., a simple parameter-estimation consistency check on the residual data) or by softening the claim and moving the one-day and white-jitter caveats into the abstract. I recommend major revision rather than rejection because the coefficient-estimation and noise-subtraction results are likely to survive such changes."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: this is a careful, well-scoped simulation study, not a new method. It asks whether the LISA TTL noise subtraction scheme still works when GW signals are present. For the five signal scenarios tested (verification binaries, SGWB, galactic binaries, MBHB mergers, and the full Sangria set), the answer is yes: coefficient errors stay below 0.1 mm/rad, residuals sit below the LISA requirement, and the injected signals look right in the ASD plots. That is genuinely useful for the LISA data-analysis community, since earlier work in this line either ignored GW signals or only touched on them.\n\nThe paper does several things well. It uses realistic simulated data (LISANode with orbit files, the LDC Sangria dataset), it checks both coefficient estimation and subtraction residuals, and it is explicit about its own limitations. The summary does not hide that the 'signal not altered' claim rests only on the indirect ASD comparisons in figures 9 and 14, for one day of data, and that longer integration and parameter estimation are left to future work. That honesty deserves credit.\n\nThe soft spots are real but not fatal. The abstract overstates: 'the GW signal characteristics were not altered' is too strong when the support is visual ASD overlap. An ASD match does not rule out phase-coherent distortion or bias in astrophysical parameters; the stress-test note is right to flag this. Second, the jitter is white in the fit band, and the authors concede they cannot say how colored jitter would change the results. Since real LISA jitter is expected to be colored, this limits direct applicability until tested. Third, the simulation uses the same linear TTL model that the fit assumes, so the test validates coefficient estimation under that model, not the model itself. That is a minor point, because the paper does not claim otherwise. Full reproducibility is also limited while the MCMC details live in an in-preparation paper.\n\nBottom line: solid engineering study with an honest discussion and one overreaching sentence in the abstract. It deserves serious peer review. I recommend accepting after the abstract is toned down and ideally after adding a parameter-estimation or colored-jitter test. I would bring it to a reading group focused on LISA data analysis.","headline":"Solid engineering validation for LISA TTL subtraction with GW signals; abstract overstates signal-preservation evidence.","tokens_in":14950,"tokens_out":3449,"would_cite":true,"duration_ms":30114,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Gravitational-wave contamination doesn't break LISA's tilt-to-length noise subtraction.","keywords":["tilt-to-length coupling","LISA","gravitational wave signals","noise subtraction","time-delay interferometry","MCMC coefficient estimation","space-based gravitational wave observatory","instrument noise modeling"],"falsifier":"Run the same one-day MCMC fit on simulated LISA data in which the spacecraft/MOSA jitter follows a colored spectrum, such as the control-loop roll-off expected in flight, with a massive black hole merger overlapping the fit band; the central claim fails if coefficient errors exceed 0.1 mm/rad or if the residual after subtracting the fitted TTL model no longer matches the injected GW waveform.","tokens_in":13964,"feed_emoji":"🛰️","tokens_out":9387,"duration_ms":80987,"temperature":0.7,"pith_summary":"LISA will measure gravitational waves from 0.1 mHz to 1 Hz, and one of its main noise sources is tilt-to-length (TTL) coupling: angular jitter of the spacecraft and optical assemblies leaking into the interferometric length readout. The mission plans to subtract this noise in post-processing by fitting TTL coupling coefficients to the data, but the same data also contain gravitational-wave (GW) signals, so the fit must not mistake signals for noise or damage them. This paper claims, using simulated LISA data, that GW responses — from verification binaries, a stellar-origin stochastic background, galactic white dwarf binaries, massive black hole mergers, and a combined multi-source data set — have little effect on the quality of the TTL fit and subtraction. In all tested cases the coupling coefficients are recovered within 0.1 mm/rad, the residual noise after subtraction falls below the LISA requirement, and the GW signal content is not perceptibly altered. If true, the planned in-flight TTL subtraction strategy can proceed without a dedicated separation of GW signals, and the science data will not be corrupted by the noise removal.","feed_headline":"GW signals don't disrupt LISA's tilt-to-length noise subtraction","feed_subtitle":"One-day simulations keep TTL coefficient errors under 0.1 mm/rad, and the gravitational waves come through unchanged.","key_machinery":"The load-bearing object is the linear tilt-to-length coupling model of Eq. (3), $\\hat{x}^{\\mathrm{TTL}}_{ij} = C^{\\varphi\\mathrm{Rx}}_{ij}\\varphi^{\\mathrm{DWS}}_{ij} + C^{\\eta\\mathrm{Rx}}_{ij}\\eta^{\\mathrm{DWS}}_{ij} + \\cdots$: each of the six links has four coupling coefficients multiplying the differential-wavefront-sensing measured angular jitter of the local and remote spacecraft/MOSA pairs, with the remote terms delayed by the light travel time. Propagated through second-generation time-delay interferometry into the orthogonal AET combinations, this model becomes the template that the Markov-chain Monte Carlo fitter matches to the simulated length data, iteratively whitening the noise between 3 mHz and 0.9 Hz; the fitted coefficients then define the subtraction. The argument's strength rests on this linear template being the true coupling and on the fit band being populated by white jitter.","core_discovery":"The central claim is that the post-processing TTL subtraction scheme planned for LISA works essentially as well when gravitational-wave signals are present as when they are absent. Using simulated one-day data with identical noise and jitter realizations, the authors fit the 24 TTL coupling coefficients of the linear model (Eq. 3) to TDI AET data with an iterative-whitening Markov-chain Monte Carlo algorithm over the 3 mHz–0.9 Hz band. For verification binaries, a stochastic gravitational-wave background, detached galactic white dwarf binaries, massive black hole binary mergers, and the combined multi-source data set, estimated coefficient deviations stay below the 0.1 mm/rad requirement, residuals after subtraction lie below the LISA mission noise requirement, and the residual TTL noise remains about an order of magnitude below the other instrument noises. Comparing the fitted-TTL-subtraction residual with the injected GW response in data that contain no other instrument noise shows that the subtraction does not perceivably alter the GW signal. The authors also show that changing arm lengths, whether from Keplerian or ESA science orbits, do not change the conclusions, supporting the static-arm assumption of earlier work.","pith_inferences":["If real spacecraft/MOSA jitter is colored rather than white, the effective TTL-dominated band shrinks and the fit leans on low frequencies where GW confusion is strongest; testing the same MCMC scheme with control-loop-shaped jitter would show whether the 0.1 mm/rad margin survives.","The one-day analysis leaves open long-integration effects: over months of data, slow drifts of coupling coefficients or a slowly evolving GW foreground could bias the TTL fit, so the claim should be re-checked on full-mission-length simulations with parameter-estimation follow-up.","The linear model in Eq. (3) omits possible nonlinear or time-varying TTL terms; a dedicated injection of a nonlinear coupling term would reveal how much unmodeled structure the fit absorbs into the linear coefficients.","Extending to source classes not tested here, such as extreme-mass-ratio inspirals or a cosmological stochastic background, would stress the low-frequency band where the GW response sits above the TTL noise; the paper itself identifies these as open cases."],"forward_implications":["The planned post-processing TTL subtraction can be run on LISA data without first separating or masking gravitational-wave signals; the tested source classes and the combined multi-source data set all leave coefficient errors below 0.1 mm/rad.","After subtraction the total noise meets the LISA requirement even when a massive black hole merger overlaps the TTL-dominated band, so the noise cleanup does not have to wait for quiet data stretches.","Because the GW response in the residuals matches the injected signal, the subtraction does not imprint a spurious waveform component, which is a prerequisite for unbiased astrophysical parameter estimation.","The static-unequal-arm assumption used in earlier TTL subtraction studies is adequate: Keplerian and ESA science orbit arm-length changes give nearly identical residuals.","The combined multi-source data set behaves like the merger-dominated case, indicating that the loudest source on a given day controls the TTL fit quality."],"supporting_citations":[{"why":"introduces the post-processing TTL subtraction scheme and the linear coupling model that this paper applies.","marker":"[7]"},{"why":"provides the iterative-whitening MCMC fitting algorithm and data-preparation details used in the coefficient fits.","marker":"[8]"},{"why":"defines second-generation time-delay interferometry combinations used to build the TDI observables.","marker":"[16]"},{"why":"gives the detailed TTL-through-TDI modeling that underlies the TTL templates used in the analysis.","marker":"[18]"},{"why":"defines the orthogonal TDI AET variables in which the coefficient fits are performed.","marker":"[19]"},{"why":"the simulator used to generate one-day interferometric data with instrument noises and jitter.","marker":"[22]"},{"why":"the package used to compute the gravitational-wave responses for the verification-binary and stochastic-background scenarios.","marker":"[27]"},{"why":"the data set used for the galactic white dwarf binary, merger, and combined multi-source scenarios.","marker":"[28]"},{"why":"provides the verification binary parameters injected as the first GW test case.","marker":"[31]"},{"why":"provides the stellar-origin binary black hole stochastic background power-law model and amplitude used in the background test.","marker":"[36]"}],"fun_headline_variants":["GW signals don't spoil LISA's noise-canceling trick","LISA tilt-to-length subtraction immune to gravitational waves","Gravitational waves fail to break LISA's TTL noise fix","Test: GWs don't mess with LISA's noise subtraction","LISA noise cleanup stays clean despite gravitational waves"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the true TTL coupling is exactly the linear model of Eq. (3) with white spacecraft and MOSA jitter in the 3 mHz–0.9 Hz fit band; if real jitter is colored or the coupling has unmodeled nonlinearity or slow drifts, the fitted coefficients and subtraction residuals could degrade beyond the reported margins.","fun_headline_variants_meta":{"raw":{"variants":["GW signals don't spoil LISA's noise-canceling trick","LISA tilt-to-length subtraction immune to gravitational waves","Gravitational waves fail to break LISA's TTL noise fix","Test: GWs don't mess with LISA's noise subtraction","LISA noise cleanup stays clean despite gravitational waves"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000231,"raw_usage":{"total_tokens":1507,"prompt_tokens":991,"completion_tokens":516,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":607,"completion_tokens_details":{"reasoning_tokens":431}},"tokens_in":607,"tokens_out":516,"duration_ms":5613,"temperature":1.0,"reasoning_tokens":431,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:25:57.646846+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same one-day MCMC fit on simulated LISA data in which the spacecraft/MOSA jitter follows a colored spectrum, such as the control-loop roll-off expected in flight, with a massive black hole merger overlapping the fit band; the central claim fails if coefficient errors exceed 0.1 mm/rad or if the residual after subtracting the fitted TTL model no longer matches the injected GW waveform.","supporting_citations":[{"cited_title":"Geometric tilt- to-length coupling in precision interferometry: mecha- nisms and analytical descriptions","cited_arxiv_id":null,"evidence_quote":"introduces the post-processing TTL subtraction scheme and the linear coupling model that this paper applies."},{"cited_title":"Paczkowski et al","cited_arxiv_id":null,"evidence_quote":"provides the iterative-whitening MCMC fitting algorithm and data-preparation details used in the coefficient fits."},{"cited_title":"Time-delay interferometry","cited_arxiv_id":null,"evidence_quote":"defines second-generation time-delay interferometry combinations used to build the TDI observables."},{"cited_title":"Auto- matic alignment of optical interferometers","cited_arxiv_id":null,"evidence_quote":"gives the detailed TTL-through-TDI modeling that underlies the TTL templates used in the analysis."},{"cited_title":"Methods for simulat- ing the readout of lengths and angles in laser interfer- ometers with Gaussian beams","cited_arxiv_id":null,"evidence_quote":"defines the orthogonal TDI AET variables in which the coefficient fits are performed."},{"cited_title":"Wanner, S","cited_arxiv_id":null,"evidence_quote":"the simulator used to generate one-day interferometric data with instrument noises and jitter."},{"cited_title":"Optical Simulations","cited_arxiv_id":null,"evidence_quote":"the package used to compute the gravitational-wave responses for the verification-binary and stochastic-background scenarios."},{"cited_title":"Unified model for the LISA measurements and instrument simulations","cited_arxiv_id":null,"evidence_quote":"the data set used for the galactic white dwarf binary, merger, and combined multi-source scenarios."},{"cited_title":"Trajectory design for the ESA LISA mission","cited_arxiv_id":null,"evidence_quote":"provides the verification binary parameters injected as the first GW test case."},{"cited_title":"LISA galactic binaries with astrometry from Gaia DR3","cited_arxiv_id":null,"evidence_quote":"provides the stellar-origin binary black hole stochastic background power-law model and amplitude used in the background test."}],"review_version":1}