{"id":"bea564e2-e700-4627-9684-b4bf8955e6b8","arxiv_id":"2506.23005","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A two-smartphone optical link at 20 cm was demonstrated; the screen's angular emission was fitted to a Lambertian source with order m=1, and success rate decays with distance.","lead":"Researchers measured how light from a smartphone screen spreads across angles and how reliably text can be sent to another smartphone camera over short distances, up to 55 cm. The result is a practical channel characterization for short-range visible light communication using ordinary screens and cameras instead of radio.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Lambertian m=1 claim rests on an unreported, uncalibrated beam-profile fit; without residuals or angular range, Eq. (1) with m=1 remains unvalidated.","rationale":"The reader's weakest-assumption analysis identifies exactly the same soft spot: the Lambertian m=1 value rests on a single normalized beam-profile fit with no angular range, residuals, or uncertainty. My stress-test agrees that this is the most load-bearing concern because the paper explicitly builds its channel model on Eq. (1) with m=1 and then uses that model implicitly in interpreting the screen-to-camera link. I considered other potential weaknesses, such as the mislabeling of A_r as the image displayed on the screen rather than the receiver area, and the lack of separate validation of the received power against distance. Those are real but secondary; the m=1 determination is the specific empirical input that the central claim depends on. The paper does not show that the screen is Lambertian over the full 0–180° range, nor that the fit is robust to receiver effects. However, this is a missing-evidence problem rather than a demonstrated contradiction. The reported success-rate roll-off with distance is qualitatively consistent with a power falloff, and there is no internal inconsistency that would force rejection. Therefore the existing CONDITIONAL verdict remains appropriate: the claim should be accepted only if the authors supply the raw beam-profile data, fit range, residuals, and a receiver-calibration check. My recommendation is UNCHANGED because the reader's verdict already makes acceptance conditional on exactly this missing support.","tokens_in":5769,"tokens_out":3024,"duration_ms":38180,"concrete_test":"Publish or recompute the fit from the Fig. 6 normalized beam-profile data over explicit angular bins (e.g., 0°, 10°, ..., 80°), fitting R(theta) = ((m+1)/(2*pi))*cos^m(theta) by nonlinear least squares, and report per-angle residuals and a 95% confidence interval for m. Also repeat the profile measurement with a second camera or with an independent photodetector to verify that the estimated m is not determined by receiver angular response. If the confidence interval excludes m=1, or if residuals exceed 10% at angles above 45°, the central Lambertian claim is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Sec. III, Fig. 6, the paper states that after normalizing the beam profile and fitting a Lambertian curve, the Lambertian order m was found to be 1. The declared FOV is 0–180°, but the paper does not state the angular range used in the fit, show fit residuals, report uncertainties, or validate the measurement setup against a known Lambertian reference. This matters because the fitted quantity cos^m(theta) is nearly flat for small angles: for theta below about 40°, m values of 0.5, 1, and 1.5 are almost indistinguishable. If the fit was performed over only the central region, the conclusion m=1 is underdetermined. Conversely, the measured beam profile is convolved with the camera's own angular response, lens falloff, and vignetting; none of these receiver effects is characterized or subtracted. Thus the estimated m=1 could be an artifact of the measurement chain rather than a property of the screen emission. This is load-bearing because the central claim is that the standard LOS channel model, Eq. (1) with m=1, describes the screen-to-camera link. The later success-rate-versus-distance result (Fig. 7) is not a check of the model: it measures decoding success rate, not received optical power, and it is presented without error bars. The claim is plausible and not contradicted by common expectations for OLED screens, but the evidence presented does not establish that Eq. (1) with m=1 is valid across the angular range and link distances used elsewhere in the paper.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper experimentally demonstrates a screen-to-camera optical camera communication (S2SVLC) link using two Google Pixel 6 Pro smartphones, reporting successful text transmission over a 20 cm link. It characterizes the smartphone screen as a Lambertian emitter with order m=1 by fitting a Lambertian curve to a measured beam profile (Sec. III, Fig. 6), and presents a success-rate versus link-span measurement from 10 to 55 cm (Fig. 7), with success rate falling to 98% at 40 cm and dropping sharply near 55 cm. The paper claims that this supports the standard line-of-sight channel model of Eq. (1) with m=1.","tokens_in":6005,"tokens_out":3747,"duration_ms":39649,"significance":"If the m=1 Lambertian characterization is reliable, it would be a practically useful result for designing screen-to-camera OCC links, because it would let link budgets use the simple cos^m(theta) model. The paper is honest that m is a fitted value rather than a theoretical prediction, and the success-rate trend is qualitatively consistent with SNR decreasing with distance. However, the evidence for the central claim is incomplete: no residuals, angular range, uncertainty, or calibration are shown for the Lambertian fit, and the success-rate curve has no error bars. The contribution is therefore an interesting experimental dataset rather than a validated channel model, and the manuscript needs substantial additional analysis to support its conclusions.","major_comments":[{"comment":"The determination of Lambertian order m=1 is not sufficiently documented. The paper does not state the angular range used in the fit, show fit residuals, report confidence intervals, or validate the measurement against a known Lambertian source. This matters because cos^m(theta) is nearly flat for theta below about 40 degrees, so fits over the central region cannot distinguish m values of 0.5, 1, or 1.5. In addition, the measured beam profile is convolved with the camera's angular response, lens falloff, and vignetting, none of which is characterized or subtracted. The authors should provide the missing fit details and ideally a residual plot, a calibrated reference measurement, and a discussion of how the receiver effects were removed; otherwise m=1 should be presented as a tentative value rather than a validated channel-model parameter.","section":"Sec. III, Fig. 6"},{"comment":"The success-rate-versus-distance measurement does not validate Eq. (1) with m=1. Success rate is a decoding metric, not received optical power, and Fig. 7 is presented without error bars, number of trials, or confidence intervals. The headline link distance of 20 cm is also not reconciled with the 10–55 cm sweep reported in the conclusion. To make the channel-characterization claim load-bearing, the authors should report repeated trials with error bars and, ideally, measure received optical power as a function of distance and angle, overlaying the predicted curve from Eq. (1) with m=1.","section":"Sec. III, Fig. 7 and Sec. IV"},{"comment":"The description of A_r as \"the image displayed in the Tx screen\" is incorrect for the standard LOS channel gain expression. In Eq. (1), A_r should denote the receiver active area (or, in an imaging context, the aperture area), not the transmitter screen area. As written, the model is dimensionally inconsistent and cannot be used for channel-gain calculations. Please correct the definition and specify how the quantities in Eq. (1) are evaluated in the experimental geometry, including the meaning of phi and psi relative to the screen and camera axes.","section":"Eq. (1) and Sec. II"}],"minor_comments":[{"comment":"The abstract states a \"link span of 20 cms\" while the success-rate sweep in Sec. III covers 10–55 cm; please clarify which link distance is the primary result and specify the conditions of the success-rate measurement.","section":"Abstract and Sec. IV"},{"comment":"The text says the angle phi is varied over 0–180 degrees, but for a flat screen the Lambertian model is physically defined over a hemisphere (0–90 degrees). Please describe the measurement geometry precisely, including how the receiver was moved relative to the screen, and avoid claiming 180-degree coverage for a cosine model.","section":"Sec. II, Fig. 3"},{"comment":"The sentence \"Note, (i) distance (d) between the Tx and the Rx to constant of 20 cm\" has grammatical errors; please revise for clarity and consistency.","section":"Sec. II, paragraph after Eq. (5)"},{"comment":"The spectral distribution in Fig. 8 is not discussed in the text; the paper would benefit from a brief analysis of the screen's RGB spectrum and its relevance to the communication channel, or the figure should be removed if it does not support a specific claim.","section":"Sec. III, Fig. 8"},{"comment":"The two beam-profiling techniques are described qualitatively, but no quantitative comparison is provided. State what the scan results show about the screen emission profile and how they relate to the Lambertian fit in Fig. 6.","section":"Sec. III, Fig. 9"},{"comment":"Reference [12] appears to concern PTP-synchronized optical switching and is not connected to the screen-to-camera content; please verify that all cited works are relevant to the claims made.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is likely a short conference-style paper; the experimental idea is reasonable, but the central Lambertian-order claim currently rests on an under-documented fit. In addition, the self-citation pattern (refs [8]–[12]) is frequent, though not disqualifying. I would advise the editor that the paper needs either a substantial expansion of the fit analysis and an independent validation of the m=1 model, or a more modest framing of the claim, before it can be accepted as a channel-characterization study."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the useful, believable part is a working 20 cm screen-to-camera text link on a Pixel 6 Pro, with a fitted Lambertian order m=1 and a success rate that decays with distance. The m=1 figure is plausible but underdocumented, and the paper never shows how this goes beyond the authors' 2023 ConTEL beam-profile paper (ref [10]). So treat it as a provisional data point, not a validated channel model.\n\nWhat they do well: the setup is honest, no tilt, dark room, real phones, and m=1 is explicitly a fitted value rather than a predicted parameter. That matters because there is no circularity in the main claim. The success-rate sweep, close to 100% at 20 cm, 98% at 40 cm, collapsing near 55 cm, is a clean and expected result. If you work on screen-to-camera OCC, this is a useful data point.\n\nThe main weakness is the Lambertian fit itself. The text says the beam profile was normalized and fitted, but it does not give the angular range used, the residuals, or any uncertainty. Over the central region, cos^m(theta) is nearly flat for m between 0.5 and 1.5, so the fit can pin m=1 only if it extends to wide angles. The camera's own angular response is not calibrated or removed, so the measured profile is a convolution of screen emission and receiver response. For the same reason, the success-rate curve does not validate Eq. (1) with m=1; it is consistent with it, but it measures decoding success, not received power. There is also a real novelty problem: ref [10] is their own earlier beam-profile paper, and this manuscript does not say what the new m=1 adds beyond it. Minor issues: the abstract says 20 cm while the sweep is 10-55 cm, and some intro sentences read like leftover drafts.\n\nIt deserves a serious referee rather than a desk reject, because the measurement is genuinely useful if the missing details can be supplied. In review, I would ask for the angular range, residuals, error bars, a check against a known Lambertian source, and a clear comparison with ref [10]. If the authors cannot separate the new result from [10], the paper should be scaled back to a short contribution that cites [10] as the earlier characterization. I would not cite m=1 as a reliable number in my own work until those details appear.","headline":"Plausible m=1 Lambertian fit and a working 20-55 cm screen-to-camera link on a Pixel 6 Pro, but missing fit details and unclear novelty over the authors' own 2023 beam-profile paper make the headline number provisional.","tokens_in":6614,"tokens_out":5444,"would_cite":false,"duration_ms":60249,"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":"The paper claims that a smartphone screen used as a transmitter behaves as a Lambertian emitter with order m=1, making the standard line-of-sight channel equation the right model for screen-to-camera links.","keywords":["optical camera communication","screen-to-camera link","Lambertian order","visible light communication","smartphone","beam profiling","line-of-sight channel","success rate"],"falsifier":"Measure the screen's radiant intensity with a goniophotometer in 1-degree steps across the full 180-degree hemisphere at constant drive level, and test the null hypothesis that R(φ) = (1/π)cos φ by computing residuals; if residuals exceed measurement noise near the edges of the field of view, the m=1 Lambertian model fails for the off-axis angles. A cheaper check: at fixed distance 20 cm, rotate the receiving phone from 0 to 80 degrees and compare the measured pixel intensity falloff to cos φ; a clear mismatch beyond the fitted range would falsify the claim.","tokens_in":5518,"feed_emoji":"📱","tokens_out":4818,"duration_ms":49566,"temperature":0.7,"pith_summary":"This paper tries to establish that a smartphone screen used as a transmitter in screen-to-camera optical camera communication behaves as a Lambertian emitter with order m=1, so the standard line-of-sight channel equation for LED links applies. The authors measured the screen's beam profile in portrait and landscape orientations, normalized it, and fitted a Lambertian curve to obtain m=1. They then demonstrated a 20 cm text link between two Google Pixel 6 Pro phones and measured the success rate of decoded text versus link span, finding 98% success at 40 cm and failure near 55 cm. The value of this claim is that S2SVLC link budgets can be computed with the same Lambertian model used for other VLC systems.","feed_headline":"Phone screens emit as Lambertian sources with order m=1","feed_subtitle":"Standard line-of-sight equations then predict screen-to-camera link power and success rate out to 55 cm.","key_machinery":"The load-bearing object is the Lambertian radiant-intensity model R(φ)=((m+1)/(2π))cos^m(φ) and its LOS channel gain H_los(0) given in Eq. (1), with m the Lambertian order. The paper parameterizes the screen as an emitter with order m and determines m by fitting the normalized measured beam profile to this one-parameter curve, obtaining m=1. The same m appears in the half-angle relation m = −ln2/ln(cos φ_{1/2}), so one measured number links the beam shape to the channel equation used for link budget and success-rate analysis.","core_discovery":"The central claim is that the emission pattern of a smartphone screen is Lambertian with order m=1, which makes the LOS DC gain H_los(0) = A_r (m+1)/(2π $d^{2}$) cos^m(φ) T_s(ψ) cos ψ with m=1 the correct channel description for a screen-to-camera link under no-tilt, no-rotation conditions. This is established by measuring the received power over 0–180 degrees in portrait and landscape configurations, normalizing the beam profile, and fitting the Lambertian curve R(φ)=(m+1)/(2π)cos^m(φ), which yields m=1. With this model the paper reports a working text link at 20 cm, success rate falling to 98% at 40 cm, and loss of link near 55 cm as received power spreads over more pixels and SNR drops.","pith_inferences":["If the Lambertian fit was limited to the central beam region, the m=1 value may not hold at large viewing angles; the paper does not report fit residuals or the fitted angular range, so a natural test is to compare the full measured profile with cos φ point-by-point.","The m=1 result implies that OLED pixel emission has a cosine angular dependence; this could be checked independently with a goniophotometer, and it suggests screen brightness and color channel may shift m with pixel content.","The same fitting method could turn any flat display into a Lambertian parameter estimate, extending VLC link models beyond OLED screens to e-ink, mini-LED, or projector screens.","A practical extension is to measure success rate versus distance for different text lengths and frame rates to see whether the 98% at 40 cm result is bit-length dependent rather than purely power dependent."],"forward_implications":["If m=1 holds, screen-to-camera link power follows cos(φ) and 1/d^2, so designers can reuse standard Lambertian LOS models for S2SVLC link budgets.","The 20 cm setup with 98% success at 40 cm gives a concrete baseline, with 55 cm as the predicted breakdown point.","Tilt and rotation, which the paper deliberately excludes, will reduce received power by cos(ψ) and cos^m(φ), so the model predicts how alignment errors hurt success rate.","The same channel characterization can be repeated for other phone models by measuring their beam profile and fitting m.","Because m=1 means a wide emission angle, a screen can serve multiple receivers at once, at the cost of lower on-axis power."],"supporting_citations":[{"why":"Supplies the Lambertian channel model in Eq. (1) and the half-angle relation in Eq. (3) that the paper fits to the screen beam profile.","marker":"[3]"},{"why":"Prior work on the same smartphone's beam profile that the paper builds on for the Lambertian fit and the frame format.","marker":"[10]"},{"why":"Defines beam profiling, the measurement concept used to obtain the screen's angular intensity distribution.","marker":"[13]"},{"why":"Describes the scanning-slit and knife-edge profiling methods the paper applies to the screen.","marker":"[14]"},{"why":"Establishes the screen-to-camera performance evaluation technique that this channel characterization extends.","marker":"[8]"}],"fun_headline_variants":["Phone screens behave as Lambertian sources with m=1","Lambertian m=1: key to screen-to-camera OCC model","Screen-to-camera link predicted up to 55 cm","Smartphone screens: Lambertian order m=1 confirmed","Channel model for screen-to-camera OCC from Lambertian fit"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"That the normalized beam profile, measured over the paper's angular range, is truly Lambertian with a single order m=1, so that Eq. (1) with m=1 holds for all angles; the paper does not report the fit's angular range, residuals, or uncertainty.","fun_headline_variants_meta":{"raw":{"variants":["Phone screens behave as Lambertian sources with m=1","Lambertian m=1: key to screen-to-camera OCC model","Screen-to-camera link predicted up to 55 cm","Smartphone screens: Lambertian order m=1 confirmed","Channel model for screen-to-camera OCC from Lambertian fit"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000224,"raw_usage":{"total_tokens":1395,"prompt_tokens":814,"completion_tokens":581,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":430,"completion_tokens_details":{"reasoning_tokens":493}},"tokens_in":430,"tokens_out":581,"duration_ms":6395,"temperature":1.0,"reasoning_tokens":493,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:52:30.632384+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the screen's radiant intensity with a goniophotometer in 1-degree steps across the full 180-degree hemisphere at constant drive level, and test the null hypothesis that R(φ) = (1/π)cos φ by computing residuals; if residuals exceed measurement noise near the edges of the field of view, the m=1 Lambertian model fails for the off-axis angles. A cheaper check: at fixed distance 20 cm, rotate the receiving phone from 0 to 80 degrees and compare the measured pixel intensity falloff to cos φ; a clear mismatch beyond the fitted range would falsify the claim.","supporting_citations":[{"cited_title":"Ghassemlooy, W","cited_arxiv_id":null,"evidence_quote":"Supplies the Lambertian channel model in Eq. (1) and the half-angle relation in Eq. (3) that the paper fits to the screen beam profile."},{"cited_title":"Beam profilers,","cited_arxiv_id":null,"evidence_quote":"Defines beam profiling, the measurement concept used to obtain the screen's angular intensity distribution."},{"cited_title":"Beam characterization: application to the laser damage threshold,","cited_arxiv_id":null,"evidence_quote":"Describes the scanning-slit and knife-edge profiling methods the paper applies to the screen."}],"review_version":1}