{"id":"d3355894-5adf-4e00-a1e3-68dc0958e2f5","arxiv_id":"1908.11023","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"LBTI's Fizeau mode is progressing toward routine phase-controlled infrared interferometry, with a PTF-slope-based correction loop tested on simulated and partially on-sky data.","lead":"This paper reports the commissioning status of the Large Binocular Telescope Interferometer's Fizeau imaging mode, including a new correction loop that uses Fourier analysis of science images to remove differential aberrations. It shares lessons from on-sky engineering tests that move the instrument toward routine high-resolution infrared imaging.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Correction algorithm validated only for noiseless, monochromatic, single-aberration simulations; simultaneous on-sky aberrations and noise may bias the PTF-slope estimate, though the paper explicitly defers this to future work.","rationale":"The reader's weakest assumption correctly identifies the simulation-to-sky transfer as the key risk, and I agree that the paper's claims are carefully limited. My refinement is that the simulation does not even exercise simultaneous OPD/tip/tilt or noise, so the PTF-slope cross-talk is untested, not merely the realism of the PSF model. This strengthens the concern slightly, but the paper is a status report: it does not claim the loop has been demonstrated on sky, and it explicitly lists the real-PSF complications as future work. Therefore the ACCEPT verdict remains appropriate; no revision of the verdict is warranted, though a cautious reader might ask the authors to report the mixed-aberration simulation or on-sky open-loop comparisons in the next revision.","tokens_in":16213,"tokens_out":11696,"duration_ms":122107,"concrete_test":"Re-run the Sec. 4.3 retrieval on simulated polychromatic Fizeau PSFs with all three aberrations varying simultaneously in a random walk, including photon noise and residual AO aberrations. If the root-mean-square error in retrieved Θ (Eq. 8) exceeds the pixel-equivalent angle PS (e.g., 10.7 mas for LMIRcam) or if the error scales with OPD, the single-DOF validation is insufficient. Alternatively, use the recorded on-sky engineering frames from fall 2018 and spring 2019: inject known PhaseCam setpoint offsets while the phase loop is closed and verify that the science-detector PTF slope predicts the resulting PSF shift within the mirror-hysteresis budget.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the Fizeau correction loop can remove NCPA rests on Eq. 8, which converts PTF slope to tip/tilt. Section 4.3 validates this on simulated PSFs with only one degree of freedom varying at a time: OPD, tip, or tilt separately, at a single wavelength (3.7 µm), diffraction-limited, with no detector or photon noise. Real Fizeau PSFs, as the authors themselves note in Sec. 4.4, contain imperfect AO correction, NCPA, ghosts, speckles, time-dependent noise, and phase smearing, and OPD, tip, and tilt occur simultaneously. The PTF slope could be biased by cross-talk between simultaneous aberrations or by the broad wavelength band and low SNR of the science detectors, causing the correction Γ = -Θ to be inaccurate and the loop to fail to converge. Since the derivation in Appendix C is heuristic (a factor-of-2 reduction for the 'center of the net illumination' is not a rigorous property of the Fizeau OTF's high-frequency lobe), the simulation is the main evidence that Eq. 8 is correct; that evidence does not yet cover the realistic regime. This is not a fatal flaw in the paper, which is a commissioning status report and explicitly states that on-sky tests of the full loop are future work, but it is the load-bearing uncertainty for the feasibility claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports the current commissioning status of LBTI's Fizeau interferometric mode. It describes the available Fizeau-Airy, non-redundant-mask, and Fizeau-grism observing modes; recent hardware upgrades (SOUL adaptive optics, OVMS+ vibration feed-forward); and the development of a correction loop that uses science-camera images to estimate differential OPD, tip, and tilt. The key algorithmic contribution is Eq. (1)/(8), which converts the slope of the phase transfer function (PTF) of the science-detector image into a wavefront tip/tilt estimate, with OPD sensed separately from a stairstep pattern in the PTF. The correction loop is validated on simulated monochromatic 3.7 micron Fizeau PSFs in which OPD, tip, or tilt performs a known random walk. The paper also reports on-sky engineering tests from fall 2018 and spring 2019, including partial phase-loop closure after the SOUL upgrade, and it closes with lessons learned and concrete next steps (capacitive mirror feedback, code porting, additional on-sky time).","tokens_in":16446,"tokens_out":8398,"duration_ms":89646,"significance":"If the proposed correction loop works on sky, it would be an important step toward routine phase-stabilized Fizeau imaging at LBTI and would provide a useful template for future ELT-scale interferometric imagers. The paper's strengths are its parameter-free derivation of the PTF-slope formula from the Fourier shift theorem, direct simulated validation with injected random walks, quantitative hysteresis measurements, and unusually candid statements of the gap between the simulated validation and the real on-sky regime. The paper does not overclaim: full closed-loop NCPA removal is explicitly deferred to future work, and the on-sky tests are presented as engineering milestones rather than as a demonstration of the complete correction loop. The main uncertainty identified by the stress-test note is real but is already acknowledged by the authors in Sec. 4.4, and it does not undermine the paper's actual status-report claim.","major_comments":[],"minor_comments":[{"comment":"The validation section shows a retrieval example for the tilt random-walk dataset, but it does not report quantitative residuals or error statistics for the OPD-only and tip-only datasets. Reporting the RMS retrieval error for all three datasets would make the simulation evidence much easier to evaluate.","section":"Sec. 4.3 / Fig. 11"},{"comment":"Equation (1) uses Nx, Ny, PS, and Delta, but those symbols are only fully defined in the appendix (Table 3). Adding a pointer to Table 3 at the first occurrence of Eq. (1), or defining the symbols inline, would improve readability.","section":"Appendix C / Sec. 4.2"},{"comment":"The sentence about the wrap-around degeneracy and the PSF elongation is vague; please state more concretely how the degeneracy is broken in the code or in post-processing.","section":"Sec. 4.3"},{"comment":"The figure legend describes colored status categories that appear only in grayscale in the printed version; using distinct symbols or hatching in addition to color would make the statuses legible in monochrome print.","section":"Fig. 4"},{"comment":"Several small typographical issues are present, e.g., 'adviseable' in the last bullet; a careful proofread of the lessons-learned section is recommended.","section":"Sec. 5"},{"comment":"The reference style 'See Fig. 1, or 8 or bottom-left panels in Fig. 11' is awkward; please rephrase as 'Fig. 1, Fig. 8, or the bottom-left panels of Fig. 11'.","section":"Sec. 2.1"}],"recommendation":"minor_revision","confidential_remarks":"This is a commissioning-status paper with a scope that fits a conference proceedings or a short instrument-report article. The authors are appropriately cautious: the correction algorithm is validated only in idealized simulations, and the full on-sky loop is explicitly future work. I see no unsupported overclaim that would merit rejection, and no novelty or citation concerns. The minor comments are presentation and completeness issues; the paper is acceptable after those are addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a well-scoped status paper, not a science result. The new pieces are the PTF-slope-based NCPA correction (Eq. 1, Appendix C) and the on-sky engineering lessons. The full correction loop has not yet been closed on sky, and the paper is upfront about that, so the main risk—that the algorithm fails on real Fizeau PSFs with simultaneous aberrations, noise, and finite bandpass—is a deferred risk, not a hidden one.\n\nWhat is genuinely new: using the phase of the science-detector OTF to estimate differential tip/tilt is a clean application of the Fourier shift theorem. The derivation in Appendix C is mostly dimensional analysis plus a factor-of-2 for the Fizeau case; that factor is plausible but heuristic, and a full OTF calculation would settle it. The simulated validation is honest: noiseless, monochromatic, one varying degree of freedom at a time. That is the right thing to do in a commissioning paper at this stage—it shows the estimator works in the ideal case without overclaiming. The MTF comparisons with Patru et al. give independent support to the sensitivity discussion. The on-sky lessons (grism focus offset, nodding behavior, phase-loop fragility after SOUL, FPC/HPC hysteresis) are practically useful; few papers publish this level of commissioning detail. The citation pattern is appropriate, and the self-citations are topical.\n\nThe soft spots are proportionate. As the stress-test notes, the algorithm has only been tested with one aberration varying at a time. On a real Fizeau PSF, OPD, tip, and tilt are present together, and the PTF slope could be biased by cross-talk or by low SNR. The authors explicitly list these complications in Sec. 4.4 and defer the full-loop test to future work, so this does not undermine the paper's stated claims. The Fig. 4 legend about code status is slightly ambiguous—some items say \"works on sky,\" others \"tested on old data\"—but that is a presentation issue, not a substantive flaw. The hysteresis measurements are empirical and worth reporting.\n\nBottom line: this paper belongs in the SPIE proceedings or a similar venue. It will be cited by anyone working on LBTI or future ELT Fizeau modes. A serious referee should check that the claims match the evidence; here they do. I would recommend accept, with minor comments asking for a two-aberration simulation (e.g., simultaneous OPD and tilt) and a cleaner statement of the code status. It is not a paradigm-shifting paper, but it is a useful and honest contribution.","headline":"A solid, honest commissioning report: the PTF-slope tip/tilt retrieval is a real new step, but the full correction loop is untested on sky and the authors say so clearly.","tokens_in":16985,"tokens_out":2272,"would_cite":true,"duration_ms":26557,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A Fourier-phase readout closes LBTI's Fizeau correction loop","keywords":["LBTI","Fizeau interferometry","non-common-path aberrations","phase transfer function","modulation transfer function","fringe tracking","infrared imaging","adaptive optics"],"falsifier":"On a bright point source, inject a sequence of known tip and tilt steps with the fast pathlength corrector mirror while the phase loop is closed, and compare the tip/tilt retrieved from the science-detector PTF slope against the commanded values; if the retrieved angles do not track the injections to within a small fraction of the plate scale, or if applying $\\vec{\\Gamma}=-\\vec{\\Theta}$ fails to hold fringe contrast on the science detector over a full exposure, the PTF-slope model is not accurate enough for closed-loop Fizeau operation.","tokens_in":1659,"feed_emoji":"🔭","tokens_out":5207,"duration_ms":82748,"temperature":0.7,"pith_summary":"The paper reports progress toward routine phase-stabilized Fizeau imaging with LBTI, the world's longest-baseline Fizeau interferometer at 22.7\\,m. It argues that a correction code reading the science camera's Fourier amplitude and phase can measure and remove the differential tip, tilt, and optical path difference that currently spoil the fringes. The central relation converts the slope of the phase transfer function into a wavefront tip-tilt vector, with the correction being its negative. If the loop works on sky, LBTI gains active removal of non-common-path aberrations and can image faint extended sources at high angular resolution in the thermal infrared; the paper presents tests on simulated PSFs and on-sky engineering data, with the full closed loop still to be demonstrated.","feed_headline":"A Fourier-phase readout closes LBTI's Fizeau loop","feed_subtitle":"PTF slopes from science frames measure tip, tilt, and OPD, enabling phase-stabilized 22.7-m infrared imaging.","key_machinery":"The central object is the phase transfer function slope formula, Eqn. 8: $\\vec{\\Theta} = [\\Omega_x N_x,\\ \\Omega_y N_y]^T (PS\\cdot\\Delta)/(\\pi\\,\\mathrm{pix}_{DFT})$, which converts the per-pixel slope of the PTF into the differential wavefront tip/tilt between the two apertures. The factor of two from the partial translation of the composite PSF and the independence from science wavelength make the formula directly usable across filters. A companion observable, the amplitude of the high-frequency lobe of the MTF, locates the center of the coherence envelope and senses OPD; a stairstep pattern in the PTF also flags OPD. The code applies the negative of the measured vector, $\\vec{\\Gamma}=-\\vec{\\Theta}$, as corrective setpoints to the phase PID loop and to internal mirrors.","core_discovery":"The paper claims that differential tip $\\Theta_y$ and tilt $\\Theta_x$ between the two LBT beams on the science detector can be read out directly from the slope $\\vec{\\Omega}$ of the phase transfer function (PTF) of the science image's Fourier transform, using $\\vec{\\Theta} = [\\Omega_x N_x,\\ \\Omega_y N_y]^T (PS\\cdot\\Delta)/(\\pi\\,\\mathrm{pix}_{DFT})$ (Eqn. 8), while optical path difference is sensed from the amplitude of the high-frequency MTF lobe. The required correction is $\\vec{\\Gamma} = -\\vec{\\Theta}$. Because the phase-sensing camera is blind to the science detector illumination in Fizeau mode, this science-detector readout closes the loop between the science focal plane and the phase-control setpoints, removing non-common-path aberrations without modifying the PhaseCam PID loop. The paper states that on-sky Fizeau engineering tests were carried out in fall 2018 and spring 2019.","pith_inferences":["The PTF-slope method is wavelength-independent in its derivation, so the same correction code could be ported between LMIRcam and NOMIC bands without re-deriving the calibration, a step the paper does not explicitly take.","If closed-loop Fizeau imaging becomes routine, LBTI's 22.7-m baseline in the thermal infrared would let it image circumstellar disks and giant-planet environments at angular resolutions comparable to future ELTs, extending the science cases the paper lists.","A testable extension would apply the same Fourier-phase readout to a single-aperture PSF to sense low-order aberrations on the science camera itself, which could complement or replace dedicated wavefront sensors in other instruments.","The correction loop's reliance on science-detector readouts means it can also serve as a fallback when the phase loop drops out mid-observation, a robustness benefit the authors mention only partially."],"forward_implications":["Fizeau observations no longer require manual alignment or \"lucky\" fringing: the code automates co-aligning the Airy PSFs, centering the coherence envelope with the grism, and closing the phase loop.","Closing the correction loop increases fringe contrast, enables longer integrations, and reduces time overheads for Fizeau science.","Phase-controlled Fizeau imaging becomes feasible for targets fainter or more extended than the current bright, point-like limit set by PhaseCam's read noise and visibility requirements.","Even with an open phase loop, the science-detector readout can partly compensate by making periodic pathlength corrections.","With capacitive position sensors installed behind the corrector mirrors, mirror commands gain closed-loop feedback, improving the reliability of both alignment and open-loop corrections."],"supporting_citations":[{"why":"Earlier description of the Fizeau non-common-path obstacles and alignment challenges that this paper's correction loop is designed to solve.","marker":"[3]"},{"why":"Predicts the MTF high-frequency lobe amplitude under AO residuals, piston errors, and phase smearing, providing the model against which empirical MTF amplitudes are compared.","marker":"[20]"},{"why":"Documents PhaseCam co-phasing status and performance, the existing phase loop that the new correction code supplements.","marker":"[21]"},{"why":"Introduces the two-band H/Ks approach to correcting phase jumps, which the correction code automates.","marker":"[24]"},{"why":"Provides the nulling-mode phase-control schematic and on-sky performance that the Fizeau loop extends.","marker":"[28]"}],"fun_headline_variants":["Science-frame slopes drive LBTI's Fizeau phase loop","PTF readout closes LBTI's Fizeau correction loop","Fizeau imaging improves with science-detector phase sensing","Phase-stabilized 22.7-m imaging from science frames"],"cache_read_input_tokens":19200,"weakest_assumption_plain":"The correction formula was validated on monochromatic, diffraction-limited simulated PSFs with only OPD, tip, and tilt as degrees of freedom, so the load-bearing premise is that the PTF-slope readout remains accurate enough on real on-sky PSFs, which include imperfect AO correction, NCPA, ghosts, speckles, detector and photon noise, and phase smearing, to close the correction loop.","fun_headline_variants_meta":{"raw":{"variants":["Science-frame slopes drive LBTI's Fizeau phase loop","PTF readout closes LBTI's Fizeau correction loop","Fizeau imaging improves with science-detector phase sensing","Phase-stabilized 22.7-m imaging from science frames"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000198,"raw_usage":{"total_tokens":1437,"prompt_tokens":1083,"completion_tokens":354,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":699,"completion_tokens_details":{"reasoning_tokens":280}},"tokens_in":699,"tokens_out":354,"duration_ms":4228,"temperature":1.0,"reasoning_tokens":280,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:28:11.672512+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On a bright point source, inject a sequence of known tip and tilt steps with the fast pathlength corrector mirror while the phase loop is closed, and compare the tip/tilt retrieved from the science-detector PTF slope against the commanded values; if the retrieved angles do not track the injections to within a small fraction of the plate scale, or if applying $\\vec{\\Gamma}=-\\vec{\\Theta}$ fails to hold fringe contrast on the science detector over a full exposure, the PTF-slope model is not accurate enough for closed-loop Fizeau operation.","supporting_citations":[{"cited_title":"Towards controlled fizeau observations with the large binocular telescope,","cited_arxiv_id":null,"evidence_quote":"Earlier description of the Fizeau non-common-path obstacles and alignment challenges that this paper's correction loop is designed to solve."},{"cited_title":"The lbti fizeau imager–ii. sensitivity of the psf and the mtf to adaptive optics errors and to piston errors,","cited_arxiv_id":null,"evidence_quote":"Predicts the MTF high-frequency lobe amplitude under AO residuals, piston errors, and phase smearing, providing the model against which empirical MTF amplitudes are compared."},{"cited_title":"Co-phasing the large binocular telescope: status and performance of lbti/phasecam,","cited_arxiv_id":null,"evidence_quote":"Documents PhaseCam co-phasing status and performance, the existing phase loop that the new correction code supplements."},{"cited_title":"A two-band approach to n λ phase error cor- rections with lbti’s phasecam,","cited_arxiv_id":null,"evidence_quote":"Introduces the two-band H/Ks approach to correcting phase jumps, which the correction code automates."},{"cited_title":"Nulling data reduction and on-sky performance of the large binocular telescope interferometer,","cited_arxiv_id":null,"evidence_quote":"Provides the nulling-mode phase-control schematic and on-sky performance that the Fizeau loop extends."}],"review_version":1}