{"id":"7ca71c93-eaeb-44db-976c-d9558dce1b7a","arxiv_id":"1908.03587","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"The VERITAS telescopes detected a spatial coherence signal from the star gamma Ori, demonstrating that Cherenkov telescope arrays can perform stellar intensity interferometry.","lead":"Astronomers report the first starlight coherence signals collected with the VERITAS gamma-ray telescopes acting as an intensity interferometer. The measurement on the star gamma Ori shows these Cherenkov telescopes can resolve stellar disks at blue wavelengths, complementing conventional optical interferometers.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Zero-lag excess may include surviving correlated noise: no null test or threshold scan is shown, so the feasibility claim rests on the assumption that the peak is purely stellar.","rationale":"The central claim is a feasibility demonstration, not a calibrated visibility measurement. The paper shows a clear zero-lag correlation excess, a plausible baseline dependence, and cumulative significance growth, all consistent with stellar intensity interferometry, and it is explicitly framed as first results. The reader's conditional verdict is appropriate: the noise-contamination caveat is serious enough that a quantitative astrophysical claim should be preceded by a null test or calibration, but it does not invalidate the demonstration. My stress-test identifies the same load-bearing assumption as the reader: the zero-lag excess is source-related rather than a surviving instrumental artifact. I do not see circularity or internal inconsistency; the HBT formalism is standard and the fits are consistency checks. The lack of machine-checked proof is not relevant for an experimental proceedings paper. No verdict change is needed.","tokens_in":7008,"tokens_out":4574,"duration_ms":56112,"concrete_test":"Re-analyze the recorded 2019-01-22/23 time series with the Fourier noise-cut threshold stepped over a factor of 2-3 (for example, 1x, 2x, and 4x the nominal excess-power threshold) while keeping all other analysis steps fixed, and plot the fitted zero-lag excess amplitude per baseline bin as a function of threshold. A pure stellar signal should remain statistically constant; an amplitude that moves monotonically with the threshold would indicate that the surviving signal is partly instrumental and would invalidate the feasibility claim as stated. This threshold scan is the decisive check because it directly tests noise contamination using already-recorded data.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing assumption is that the zero-lag correlation excess in Figure 2 is caused by the spatial coherence of gamma Ori rather than by residual instrumentally correlated noise. The paper's own data-quality discussion states that high-frequency noise renders 30-50% of the one-second correlograms unusable and that an exact solution is currently unknown. The Fourier-based cut rejects cycles with excess power at the identified noise frequency, but it does not demonstrate that surviving cycles are free of broadband, intermittent, or common-mode correlations. Because the optical-path-delay correction places the expected HBT signal at the zero-lag bin, any correlated electronic pickup with the same relative time delay would survive at exactly the position where the claim is made. The right-panel SNR proportional to sqrt(t) growth is not discriminating: a stable spurious excess also accumulates as sqrt(t). The left-panel baseline dependence is the best evidence for a stellar origin, but the fits use free normalization and peak lag, and the paper does not compare the measured excess amplitudes with the squared visibility predicted from theta_UD = 0.701 mas or with a zero-baseline or calibrator observation. The feasibility claim is therefore not fully secured until residual noise contamination is excluded.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the first stellar intensity interferometry (SII) observations with two VERITAS IACT telescopes, targeting the B-type star gamma Ori on 2019 January 22. After applying time-delay corrections and a Fourier-based data-quality cut, the authors average one-second correlograms and observe an excess at zero lag that they attribute to spatial coherence. The excess amplitude increases as the projected baseline decreases, and the cumulative significance grows approximately as the square root of integration time. The paper interprets these results as demonstrating the feasibility of SII with IACT arrays and the first 'off-line' optical interferometer, with data streamed to disk and correlated post-observation. The instrumentation, observation strategy, and the ultraviolet-optical complementarity to amplitude interferometers are described in the context of the future Cherenkov Telescope Array.","tokens_in":7161,"tokens_out":3067,"duration_ms":35631,"significance":"If the reported zero-lag excess is genuinely stellar, this is a valuable proof-of-feasibility result: it would show that existing IACT arrays can be retrofitted for optical intensity interferometry, that observations can be conducted in bright-moon time without impacting gamma-ray programs, and that offline correlation of streamed data is practical. The paper is also useful in outlining a path toward CTA-era SII. The authors correctly apply the standard Hanbury Brown-Twiss formalism of Eq. (2.1)-(2.2), and they are candid about the data-quality losses and the current lack of an exact solution for the noise. However, the quantitative evidence presented is incomplete: the measured excess is not placed on an absolute visibility scale, no statistical uncertainties are shown on the binned correlation curves, and no comparison is made to the published uniform-disk diameter of gamma Ori. The feasibility conclusion is therefore plausible but not yet fully secured by the analysis as written.","major_comments":[{"comment":"The paper never provides an absolute calibration of the g(2) excess amplitude, so it cannot verify Eq. (2.1) or compare the measured squared visibility with the value predicted from the published uniform-disk diameter theta_UD = 0.701 ± 0.005 mas. The sinc fit in Figure 2 uses a free normalization and a free peak time-lag, which means the amplitude is not used as a quantitative test. For a claim of 'astrophysical measurements', the authors should report the measured correlation amplitude in physical units and compare it with the expected |V(r)|^2 from Eq. (2.2), including uncertainties.","section":"Section 2, Eq. (2.1) and Figure 2"},{"comment":"The paper states that 30-50% of the one-second correlograms are discarded because of high-frequency noise and that 'an exact solution is currently unknown'. The Fourier-based cut rejects cycles with excess power at the identified noise frequency, but it does not demonstrate that the surviving cycles are free of broadband, intermittent, or common-mode correlations. Because the optical-path-delay correction places the expected HBT signal at the zero-lag bin, any residual correlated electronic pickup with the same relative delay would appear at exactly the position where the claim is made. A null test (e.g., an off-source observation, a calibrator star, or a noise-threshold scan) is needed to exclude this possibility.","section":"Section 2, data quality cut paragraph"},{"comment":"The binned correlation curves in Figure 2 are shown without error bars or per-bin significance estimates. As a result, it is not possible to assess whether the apparent increase of the excess at shorter baselines is statistically significant or to quantify the precision of the measurement. The authors should provide uncertainties on the binned correlation values and, ideally, overplot the predicted visibility curve for the published gamma Ori diameter.","section":"Figure 2, left panel"},{"comment":"The cumulative SNR growth fit of the form SNR = A*sqrt(t), with A as a free parameter, is not a discriminating test of a stellar origin for the excess. A stable spurious correlation would also accumulate as sqrt(t). To support the stellar interpretation, the authors should show that the normalized excess amplitude decreases with increasing baseline in a way consistent with the expected visibility, or demonstrate the absence of a similar excess in a control observation where no spatial coherence is expected.","section":"Figure 2, right panel"}],"minor_comments":[{"comment":"The notation g(2)(tau) should be typeset as g^(2)(tau) with the superscript, and the text around Eq. (2.1) would benefit from a brief definition of the angle brackets as time averages.","section":"Equation (2.1)"},{"comment":"The sentence 'Note more negative values correspond to higher current' is confusing; please state the sign convention explicitly, e.g., that the plotted quantity is the negative of the PMT current.","section":"Figure 1 caption"},{"comment":"Reference [23] is incomplete: it cites 'in proceedings of 36th International Cosmic Ray Conference (these proceedings)' without a paper title or authors; also, the spelling of Hanbury Brown is inconsistent between references [1]-[3] and [26].","section":"References"},{"comment":"The statement that the system 'has already been extended to the two additional telescopes' appears only in the abstract and conclusions; the body of the paper describes only the T3/T4 pair, so a sentence in Section 2 describing the extension would improve consistency.","section":"Section 3, Conclusions"}],"recommendation":"major_revision","confidential_remarks":"This is a proceedings-style contribution, and the bar for 'demonstrating feasibility' may be lower than for a full research article. However, the current text overstates the strength of the evidence: the zero-lag excess, the baseline dependence, and the SNR growth are suggestive but not quantitatively established without an absolute visibility calibration or a noise-exclusion null test. These are fixable within the manuscript's scope, so I recommend major revision rather than rejection. The paper would also benefit from a statement clarifying whether the 'first results' claim refers only to the VERITAS system or to IACT-based SII more broadly, given the prior work cited in references [6] and [7]."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe claim that matters is that IACT arrays can do stellar intensity interferometry, and this paper makes a credible case. The demonstration with two VERITAS telescopes at 81.5 m baseline using an offline correlator is new; Guerin et al. did spatial SII on stars with a 15 m baseline and much smaller collectors. The baseline-dependent rise in the correlation excess is a genuinely physical signature, and the authors are honest that 30–50% of the data is corrupted by high-frequency noise and that the exact remedy is unknown.\n\nThe paper does not oversell itself. It is framed as a feasibility demonstration, which it likely is. Standard HBT theory is used without circular assumptions; the fits are consistency checks. The engineering is sound and the path to all four telescopes is sensible.\n\nThe soft spots are the ones you would expect: no absolute visibility, no error bars on the binned correlation curves, no quantitative comparison to the published 0.701 mas diameter, and no null test or calibrator. The stress-test note is right that a broadband correlated pickup at the optical path delay would land at the same zero-lag bin, and a stable spurious excess would also accumulate as sqrt(t). So the central claim rests partly on the assumption that the peak is stellar. That is a real weakness, but it is not fatal for a proceedings paper whose purpose is to show the system works. The baseline dependence makes the peak much more likely to be real than not.\n\nWhat is missing is a calibration or a null observation on a target expected to show no correlation, and a direct comparison of measured amplitude to the visibility predicted from the known diameter. Without those, the quantitative astrophysics is not yet secure. For a feasibility claim, this is adequate.\n\nWho should read it: anyone planning SII with CTA or other IACT arrays, and people trying to build offline correlators. It deserves a serious referee if expanded into a full instrument paper. I would like to see the missing calibration data before citing it as a measurement.\n\nRecommendation: engage with it, but treat it as an engineering milestone, not an astrophysical result.","headline":"A credible feasibility demonstration of SII with IACT arrays that deserves engagement, but the quantitative case needs a null test and a measured visibility before it becomes an astrophysical result.","tokens_in":7748,"tokens_out":2015,"would_cite":true,"duration_ms":21526,"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":"The VERITAS Stellar Intensity Interferometer detects spatial coherence in starlight from the star γ Ori, demonstrating that Imaging Atmospheric Cherenkov Telescope arrays can perform long-baseline optical intensity interferometry.","keywords":["stellar intensity interferometry","VERITAS","Imaging Atmospheric Cherenkov Telescope","spatial coherence","squared visibility","off-line interferometry","Cherenkov telescope array"],"falsifier":"Point the two telescopes at an empty patch of sky (or at a star known to be unresolved at this baseline) and apply the same analysis; if the zero-lag excess persists, the signal is an instrumental artifact rather than stellar coherence.","tokens_in":1642,"feed_emoji":"🔭","tokens_out":2453,"duration_ms":67126,"temperature":0.7,"pith_summary":"This paper reports the first astrophysical measurements made with the VERITAS Stellar Intensity Interferometer, an instrument that turns four 12-meter gamma-ray telescopes into an optical intensity interferometer. Observations of the bright star γ Ori (Bellatrix) using two telescopes show a clear correlation excess at zero time-lag, the expected signature of spatial coherence, with the amplitude growing as the projected telescope baseline shrinks, matching the Airy-disk visibility profile of a resolved uniform disk. The cumulative signal-to-noise ratio grows as the square root of integration time, confirming that the system behaves predictably. The authors conclude that this demonstrates the feasibility of Stellar Intensity Interferometry with Imaging Atmospheric Cherenkov Telescope arrays and introduces the first 'off-line' optical interferometer, where data are streamed to disk and correlated post-observation.","feed_headline":"Starlight coherence measured with gamma-ray telescopes","feed_subtitle":"Two VERITAS telescopes detect Bellatrix's spatial coherence, opening a new path to stellar diameters.","key_machinery":"The central relation is the Hanbury Brown-Twiss intensity-interferometry formula $g^{(2)}(r,\\tau=0) = 1 + (\\Delta f/\\Delta\\nu)|V(r)|^2$, connecting the zero-lag intensity correlation to the squared visibility of the source, with $V(r)$ given by the Airy-disk profile $V(r)=2J_1(\\pi\\theta r/\\lambda)/(\\pi\\theta r/\\lambda)$ for a uniform-disk star. The instrument streams photomultiplier currents from each telescope, applies a time-delay correction for the optical path difference and cable delays, discards one-second cycles contaminated by narrow-band high-frequency noise (identified by Fourier analysis), and computes correlations offline with an FPGA correlator. The measured correlation peak is fit to $g^{(2)} \\propto \\operatorname{sinc}(\\pi\\Delta f\\tau)$, and the baseline is varied naturally by tracking the star across the sky (uv-plane synthesis).","core_discovery":"The central claim is that a long-baseline optical intensity interferometer can be built from existing gamma-ray telescopes with modest instrumentation. Using two VERITAS telescopes separated by a radial baseline of about 81.5 meters, the system measured the second-order coherence function $g^{(2)}(\\tau)$ of the B-type star γ Ori at an effective wavelength of 415 nm. The measured zero-lag correlation excess follows the expected sinc profile for a rectangular electronic bandwidth, increases as the projected baseline decreases (consistent with the visibility of a uniform disk of angular diameter ~0.7 mas), and accumulates significance as SNR $\\propto \\sqrt{t}$. This establishes that IACT arrays, despite their relatively poor optical quality, can collect usable intensity-interferometry data, reaching stars about three magnitudes fainter than the historical Narrabri Stellar Intensity Interferometer and complementing amplitude interferometers at shorter wavelengths.","pith_inferences":["If the zero-lag excess is indeed stellar, re-analyzing the discarded 30-50% of noise-affected cycles with improved noise-removal algorithms could boost sensitivity without any hardware change.","The technique's limiting magnitude is set by the telescope point-spread function and night-sky background; next-generation Cherenkov telescopes with better PSFs (such as Schwarzschild-Couder designs) should extend SII to fainter targets.","Other existing IACT arrays, such as MAGIC and HESS, could be retrofitted with similar SII instrumentation, effectively creating a global network of optical intensity interferometers with baselines of hundreds of meters.","Because the stored intensity streams can be re-correlated with updated time-delay or noise models, the data constitute an archival resource for 'virtual interferometry' long after the observation night."],"forward_implications":["The same two-telescope setup can measure the angular diameters of bright stars by fitting the squared visibility as a function of projected baseline.","Because correlations are computed offline, SII observations can be scheduled during bright-moon periods without affecting the primary gamma-ray observing program.","Doubling the sampling rate from 250 MS/s to 500 MS/s should improve the signal-to-noise ratio by about $\\sqrt{2}$, as stated in the paper.","Adding the other two VERITAS telescopes yields six baselines and denser uv-plane coverage, enabling measurements of stellar shapes such as the oblateness of rapidly rotating stars.","The same instrumentation approach could be applied to future Cherenkov arrays like CTA, which could achieve a limiting magnitude about three magnitudes fainter than the Narrabri interferometer."],"supporting_citations":[{"why":"Supplies the basic theory that intensity fluctuations in two separated beams are correlated according to the source spatial coherence.","marker":"[1]"},{"why":"Provides the experimental test of the theory for partially coherent light, establishing the foundation of intensity interferometry.","marker":"[2]"},{"why":"Reports the Narrabri Stellar Intensity Interferometer measurements of 32 stellar diameters, the historical precedent this work extends.","marker":"[3]"},{"why":"Shows quantitatively that Imaging Atmospheric Cherenkov Telescope arrays are well-suited for optical intensity interferometry.","marker":"[8]"},{"why":"Describes the laboratory digital intensity interferometer with thermal light that the VERITAS hardware is based on.","marker":"[21]"},{"why":"Provides the measured uniform-disk angular diameter of γ Ori used to select the target and to compare with the visibility results.","marker":"[25]"},{"why":"Gives the sinc-function form for the correlation peak under a rectangular electronic bandwidth, used to fit the measured correlogram.","marker":"[26]"}],"fun_headline_variants":["Gamma-ray scopes pin down Bellatrix's size","Gamma-ray telescopes measure stellar angular size","VERITAS scopes double as stellar interferometer","First long-baseline stellar coherence with IACTs","Bellatrix's angular size from gamma-ray scopes"],"cache_read_input_tokens":9856,"weakest_assumption_plain":"The zero-lag correlation excess attributed to stellar spatial coherence is assumed to come from the star and not from residual instrumental noise that survives the Fourier-based data-quality cut.","fun_headline_variants_meta":{"raw":{"variants":["Gamma-ray scopes pin down Bellatrix's size","Gamma-ray telescopes measure stellar angular size","VERITAS scopes double as stellar interferometer","First long-baseline stellar coherence with IACTs","Bellatrix's angular size from gamma-ray scopes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000784,"raw_usage":{"total_tokens":3410,"prompt_tokens":843,"completion_tokens":2567,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":459,"completion_tokens_details":{"reasoning_tokens":2493}},"tokens_in":459,"tokens_out":2567,"duration_ms":17865,"temperature":1.0,"reasoning_tokens":2493,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:08:15.176454+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Point the two telescopes at an empty patch of sky (or at a star known to be unresolved at this baseline) and apply the same analysis; if the zero-lag excess persists, the signal is an instrumental artifact rather than stellar coherence.","supporting_citations":[{"cited_title":"Hanbury Brown and R","cited_arxiv_id":null,"evidence_quote":"Supplies the basic theory that intensity fluctuations in two separated beams are correlated according to the source spatial coherence."},{"cited_title":"Hanbury Brown and R","cited_arxiv_id":null,"evidence_quote":"Provides the experimental test of the theory for partially coherent light, establishing the foundation of intensity interferometry."},{"cited_title":"Hanbury Brown, J","cited_arxiv_id":null,"evidence_quote":"Reports the Narrabri Stellar Intensity Interferometer measurements of 32 stellar diameters, the historical precedent this work extends."},{"cited_title":"LeBohec and J","cited_arxiv_id":null,"evidence_quote":"Shows quantitatively that Imaging Atmospheric Cherenkov Telescope arrays are well-suited for optical intensity interferometry."},{"cited_title":"Matthews, D","cited_arxiv_id":null,"evidence_quote":"Describes the laboratory digital intensity interferometer with thermal light that the VERITAS hardware is based on."},{"cited_title":"Hanbury-Brown, The Intensity Interferometer","cited_arxiv_id":null,"evidence_quote":"Gives the sinc-function form for the correlation peak under a rectangular electronic bandwidth, used to fit the measured correlogram."}],"review_version":1}