{"id":"2463ca68-7580-4d35-ac98-02ee5499d148","arxiv_id":"2412.07359","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"The measured path gain through 304 GHz static RISs matches a near-field radar model only after a 5.5 dB offset is applied to all data.","lead":"The paper reports indoor channel measurements at 304 GHz using three static reconfigurable intelligent surface (RIS) prototypes with 1-, 2-, and 3-bit phase quantization, and compares them with analytical models. A smart generalist might read it to see whether current near-field models for THz RIS-assisted links hold up in real rooms, and how much the measurements had to be adjusted to fit.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'excellent agreement' that validates Eq. (2) is established only after a 5.5 dB correction taken from an unreviewed companion paper [12]; without an independent calibration trace this is not a test of the near-field model.","rationale":"I agree with the reader's weakest-assumption analysis. The paper's strongest claim is the quantitative validation of the near-/far-field path-gain formula (2), and the only evidence for that claim is Fig. 4, where the measured data have been shifted by a 5.5 dB constant. Because that constant is sourced from [12], which is under review and not included, the validation is not independently testable from this manuscript. The concern is load-bearing: if the offset is a fitted nuisance parameter, the absolute-level match is meaningless, and even the shape match cannot be assessed without residuals and uncertainties. I do not see the concern as fatal—the measurement campaign appears carefully executed, and the near-field trend in Fig. 4 is qualitatively plausible—but it does justify a conditional acceptance requiring disclosure of the calibration procedure and raw data. The reader already reached this verdict, so I recommend leaving it unchanged. A useful additional check would be to re-derive Eq. (2) from [9, Lemma 1] to confirm that the K factor is appropriate for a bi-static RCS formulation with a 30-degree tilted beam and quantized phase; however, the immediate blocking issue is the calibration offset.","tokens_in":9247,"tokens_out":5777,"duration_ms":61147,"concrete_test":"Obtain from the authors the raw measured path gains before the 5.5 dB correction, together with the calibration measurement that produced the 5.5 dB value in [12]. Fit a single additive constant to the raw data against Eq. (2) over d2 >= 0.25 m and compare the best-fit constant with 5.5 dB; also compute per-distance residuals. If the best-fit offset differs from 5.5 dB by more than the stated measurement uncertainty, or if the residuals show a systematic d2 trend, the claimed agreement is not independently supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is in Section III-B: measured path gains are shifted by a constant 5.5 dB before comparison with formulas (1) and (2), with the offset attributed to [12], a companion paper under review. The central claim that Eq. (2) is validated by 'excellent agreement' for d2 >= 0.25 m is therefore conditioned on an offset whose magnitude and validity are not independently established in this manuscript. If the 5.5 dB value was selected to align measured and modeled curves, the agreement in absolute level is uninformative; if it was transferred from a different measurement campaign, transferability to this setup has not been shown. A constant dB shift also cannot remove distance-dependent systematic errors such as misalignment or beam squint, which are most relevant in the near-field region where (2) is claimed to improve over (1). Combined with the absence of error bars and raw data, the agreement in Fig. 4 cannot be quantitatively assessed. This does not invalidate the measurement campaign or the design contributions, but it does mean the central validation claim currently rests on an unverified calibration constant.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports an indoor measurement campaign at 304 GHz using three static RISs with 1-, 2-, and 3-bit phase quantization, each with 100x100 half-wavelength unit cells designed for a non-specular reflection at 30 degrees. In Setup 1, the measured Tx-RIS-Rx path gain is compared with the far-field radar equation (1) and with a near-/far-field corrected formula (2) that uses the RCS values of Table I and the correction factor K from reference [9]; after applying a constant 5.5 dB correction attributed to reference [12], the paper claims excellent agreement for d2 >= 0.25 m and concludes that the near-field analysis in [9] is validated. In Setup 2, PAP measurements with the 1-bit RIS, the 3-bit RIS, and a PEC backflip are compared qualitatively, and the paper closes with a list of challenges and research directions for THz RIS design and channel modeling.","tokens_in":9465,"tokens_out":5311,"duration_ms":57081,"significance":"If the validation claim holds, this would be one of the first experimental confirmations of a near-field beamforming-error model for RIS-assisted links at sub-THz frequencies. The paper has clear strengths: the three hardware prototypes are described in detail, the RCS values and aperture efficiencies are reported, full-wave simulations are used to support the design, and the PAP comparison against a PEC baseline is a useful experimental reference. The main limitation is that the quantitative validation of Eq. (2) depends on a 5.5 dB constant correction taken from an under-review companion paper, with no error bars or uncorrected data shown. Because the central claim rests on this calibration step, the paper is not yet a self-contained quantitative validation, although the measurement campaign and design contributions remain valuable.","major_comments":[{"comment":"The central validation is conditional on the 5.5 dB correction from reference [12]. This correction is not derived from an independent calibration in the present manuscript, the uncorrected measurements are not shown, and no error bars or noise floor information are provided. Since the claimed 'excellent agreement' concerns absolute path-gain levels, the reader cannot test whether Eq. (2) is actually validated. Please show the raw (uncorrected) data, quantify measurement uncertainty, and provide a sensitivity analysis with respect to the offset value (e.g., for offsets in the range 3-8 dB).","section":"Section III-B, Eq. (2), Fig. 4"},{"comment":"A constant dB shift cannot absorb distance-dependent systematic errors such as pointing misalignment, beam squint, or near-field multipath contributions. The distinguishing feature of the data is the growing gap between Eq. (1) and measurements at small d2; if any of these errors also depends on d2, the constant offset could either create or mask the near-field effect attributed to K. Please provide evidence that the 5.5 dB offset is constant over the entire d2 range and quantify the residual distance-dependent uncertainty before claiming that the near-field formula is validated.","section":"Section III-B, Fig. 4"},{"comment":"The near-field correction K is taken from reference [9], authored by two co-authors of this manuscript, and the 5.5 dB offset is taken from reference [12], with overlapping authorship. This does not by itself invalidate the comparison, but it makes the validation of Eq. (2) partially self-referential. A stronger test would be a comparison against an independently computed full-wave prediction of the Tx-RIS-Rx path gain, or at least a quantitative uncertainty budget for K and for the offset. Please add such a benchmark or an explicit statement of the independence of the correction.","section":"Section III-B and Refs. [9], [12]"},{"comment":"The PAP analysis is qualitative. In particular, the claim that the 3-bit RIS 'results in larger spatial diversity' than the 1-bit RIS is not supported by a quantitative metric. If the PAP comparison is intended as a characterization result, please report angular spread values, the number of MPCs above a defined threshold, and the associated measurement uncertainty, rather than relying only on visual inspection of Fig. 5.","section":"Section III-C, Fig. 5"}],"minor_comments":[{"comment":"The text says the simulated radiation patterns are almost identical for r > 4 m, but later states they are almost identical for r > 1 m; please reconcile these statements and indicate the applicable distance on Fig. 2.","section":"Section II, Fig. 2"},{"comment":"The x-axis tick labels appear corrupted or duplicated (e.g., '0 0.2 0.4 1 1 1.5 2'); please correct the axis and consider marking the key distances d2 = 1.15 m and d2 = 0.25 m explicitly.","section":"Fig. 4"},{"comment":"Several references directly used in the quantitative analysis are under review, and no preprint is given for [12], which provides the 5.5 dB correction. Since the reader cannot verify the correction, please include a public preprint or an appendix documenting the measurement conditions and calibration methodology of [12].","section":"References [9], [11], [12]"},{"comment":"The PAP subfigures use circle sizes to indicate relative MPC strength, but the absolute normalization and the detection threshold for what counts as an MPC are not stated; please specify them so the comparison between the RIS and PEC cases is reproducible.","section":"Section III-C, Fig. 5"},{"comment":"The abstract uses 'field-trial measurements' while the experiments are conducted in a lecture room; 'indoor measurement campaign' would be more precise. Also, in Section IV-B the statement that the PEC yields 'more channel paths' should be quantified with the MPC detection threshold just mentioned.","section":"Abstract and Section IV-B"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's quantitative validation rests on two under-review companion papers [9] and [12] with overlapping authorship. The editors may wish to verify that the 5.5 dB correction in [12] is transferable to the setup reported here and that the near-field model in [9] has an independent basis. The paper has substantial experimental value, but the central comparison should be made self-contained before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nRead the 304 GHz RIS measurement paper. The genuinely new thing is the experimental campaign: three static RIS prototypes with 1/2/3-bit phase quantization, 100x100 elements, measured at 304 GHz with a real channel sounder. That data is not in the prior literature, and the fabrication/characterization section (RCS values, aperture efficiencies, full-wave simulation comparison) is done carefully. The PAP measurements in the blocked-LoS setup are also useful, though that analysis is qualitative.\n\nThe soft spot is exactly where the reader's report puts it. The central validation of Eq. (2) — the near-field corrected path gain — depends on shifting all measured gains by a constant 5.5 dB, and that offset comes from a companion paper under review [12]. If that offset is a fitted or campaign-specific calibration, the 'excellent agreement' in Fig. 4 is not an independent test. If it was transferred from a different setup, the transferability has not been shown. Either way, the paper does not give error bars or raw data, so the agreement can only be judged by eye. Also, the near-field correction K is taken from [9], by two of the co-authors, so the validation is partly self-referential: own model, own measurements, own correction. That is not a fatal circularity — the RCS values are independent and the shape of the d2 dependence is meaningful — but it lowers the strength of the claim.\n\nI would not call the central conclusion false. The distance-dependent behavior — far-field formula overestimates at short d2, the near-field correction tracks the measurements — is visible in the figure and consistent with the full-wave simulations in Section II. The 5.5 dB offset affects the absolute level, not the shape. So the qualitative finding (near-field matters below ~1.15 m) probably stands. But the quantitative validation of Eq. (2) is conditional on disclosing the calibration procedure.\n\nRecommendation: send it to review, but require the authors to report the calibration trace for the 5.5 dB offset, provide error bars or raw data, and ideally compare against an independent model or a direct calibration measurement. The paper is worth a serious referee because the dataset is new and the prototypes are real. Just do not let the 'excellent agreement' claim through without those numbers.","headline":"New 304 GHz RIS measurements are valuable, but the headline near-field validation rests on a 5.5 dB offset from an unreviewed companion paper, so treat the quantitative agreement as conditional.","tokens_in":10047,"tokens_out":1814,"would_cite":true,"duration_ms":17543,"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 paper claims that a closed-form near-field correction predicts measured 304 GHz RIS path gains down to 0.25 m, after a constant 5.5 dB systematic offset is applied.","keywords":["reconfigurable intelligent surfaces","THz communications","304 GHz channel sounding","path loss modeling","near-field beamforming","power angular profile","indoor propagation"],"falsifier":"Run the same setup with an independent absolute calibration of the 304 GHz channel sounder (for example, a direct cable-back-to-back measurement or a known free-space path between two horn antennas) and re-measure the Tx-RIS-Rx path gain without applying any offset; if the measured curve between $d_2 = 0.25$ m and $1.15$ m deviates from formula (2) by more than the reported 2.8 dB, the validation collapses.","tokens_in":9074,"feed_emoji":"📡","tokens_out":8224,"duration_ms":70076,"temperature":0.7,"pith_summary":"This paper reports indoor channel measurements at 304 GHz in which the transmitter–receiver link is completed through a reconfigurable intelligent surface (RIS), using three static RIS prototypes with 1-, 2-, and 3-bit phase quantization. Its central claim is that the measured Tx-RIS-Rx path gain follows the near-/far-field analytical formula (2), which multiplies the standard bistatic radar equation by a beamforming-error factor $K$, with close agreement for RIS-receiver distances down to 0.25 m. The agreement is obtained after a constant 5.5 dB offset is applied to the measured data to absorb systematic errors such as sounder instability, misalignment, and fabrication imperfections. The authors argue this validates a recently proposed near-field beamforming model for sub-THz RIS links and shows that a 3-bit RIS reshapes the indoor multipath profile, reducing leakage compared with a flat reflector. The significance is that system designers could rely on a simple closed-form path-gain model instead of full-wave simulation or ray tracing when planning indoor THz coverage assisted by RISs.","feed_headline":"Measured 304 GHz RIS link matches near-field model to 0.25 m","feed_subtitle":"Indoor experiments with 1- to 3-bit RIS prototypes confirm a closed-form beamforming-error correction.","key_machinery":"The central object is the near-/far-field path-gain formula (2), $P_{\\mathrm{nf,ff}} = K^2(d_2,30^\\circ)\\,P_{\\mathrm{ff}}$, which multiplies the conventional bistatic-radar path gain by the squared beamforming-error factor $K(d_2,30^\\circ)$ taken from Lemma 1 of the authors' earlier work. $K$ quantifies the gain loss when the receiver at distance $d_2$ and angle $30^\\circ$ sits inside the depth of focus of the RIS rather than in its far field. The argument is carried by three fabricated static RISs with $100\\times100$ half-wavelength-spaced unit cells and 1-, 2-, and 3-bit phase quantization, whose measured RCS values (Table I) provide the $\\sigma_{\\mathrm{RIS}}$ entering the far-field formula. The $K$ correction is the load-bearing piece: it converts a far-field model that overestimates short-distance gain into one that tracks the measurements down to $0.25$ m.","core_discovery":"The paper's central claim is that the end-to-end gain of an indoor Tx-RIS-Rx link at 304 GHz, measured over RIS-receiver distances from 0.2 to 10 m, is predicted by the near-/far-field formula $P_{\\mathrm{nf,ff}} = K^2(d_2,30^\\circ)\\,P_{\\mathrm{ff}}$, where $P_{\\mathrm{ff}} = \\sigma_{\\mathrm{RIS}}\\lambda^2/((4\\pi)^3 d_1^2 d_2^2)$ is the standard bistatic radar equation built from measured RCS values, and $K$ is a near-field beamforming-error factor computed from the RIS's effective aperture. The authors report that, once a constant 5.5 dB correction is applied to the measurements to absorb systematic errors, the formula matches the data for $d_2 \\geq 0.25$ m, whereas the far-field-only formula overestimates the gain at short distances; at $d_2 = 0.41$ m the measured deviation from the far-field prediction is about 2.8 dB, close to the predicted 3 dB. The paper further claims that power angular profile measurements show the 3-bit RIS concentrating power in its designed $30^\\circ$ beam and reducing scattered multipath relative to a flat conducting reflector, which it reads as evidence that phase-quantized RISs can shape indoor sub-THz channels.","pith_inferences":["If the 5.5 dB offset is confirmed by an independent calibration, the same $K$-correction should scale with effective aperture to other sub-THz RIS sizes and frequencies, giving a parameter-light way to predict short-range RIS gains.","The angular-profile data suggest a design trade-off: a flat conducting surface increases spatial multipath diversity, while a focused RIS suppresses alternative paths; which one helps depends on whether the goal is spatial multiplexing or interference control.","A sharper test of the near-field model would be a full angular sweep at a fixed $d_2$, comparing measured and simulated patterns in both angle and distance, since the current agreement is demonstrated mainly along the designed $30^\\circ$ cut.","Independent replication with a calibrated sounder and a different RIS fabrication run would settle whether the 5.5 dB correction and the near-field factor $K$ are general or specific to this prototype set."],"forward_implications":["Indoor THz link budgets can use the closed-form formula (2) with effective aperture instead of full-wave simulation for RIS-receiver distances above about 0.25 m.","Below the depth-of-focus distance $r_{\\mathrm{DF}} \\approx 1.15$ m (refined to 0.41 m with effective aperture), the far-field formula (1) overestimates the path gain, so near-field-aware models are needed for short-range RIS deployments.","Higher phase resolution yields sharper beam shaping: the 3-bit RIS shows a 15 dB ratio between main and mirror beams and produces the strongest measured multipath component, while the 1-bit RIS splits power equally and lets a wall reflection dominate.","An optimized non-specular RIS reduces unwanted scattering compared with a flat conducting reflector, so the RIS phase profile can shape the indoor angular spread rather than merely adding reflections."],"supporting_citations":[{"why":"Supplies the near-field beamforming-error factor K in formula (2), which is the main validation target.","marker":"[9]"},{"why":"Provides the 5.5 dB systematic correction applied to all measured path gains before comparison.","marker":"[12]"},{"why":"Describes the 304 GHz channel sounder and its measurement methodology.","marker":"[10]"},{"why":"Defines the depth-of-focus distance used to interpret where far-field predictions break down.","marker":"[8]"},{"why":"Defines Fresnel and Rayleigh distances that locate the RIS near-field region.","marker":"[7]"},{"why":"Underlies the double-directional power angular profile measurement approach in Setup 2.","marker":"[11]"}],"fun_headline_variants":["304 GHz RIS link matches near-field model from 0.25 m","Indoor RIS measurements at 304 GHz validate near-field gain formula","Near-field correction makes 304 GHz RIS gain model fit measurements","Phase-quantized RIS at 304 GHz fits near-field model from 0.25 m","Indoor 304 GHz RIS tests confirm near-field gain correction model"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a single constant 5.5 dB correction, taken from a separate measurement study rather than an independent calibration of this setup, fully accounts for all systematic errors in the measured path gains; without that offset, the claimed agreement between the data and formula (2) does not hold.","fun_headline_variants_meta":{"raw":{"variants":["304 GHz RIS link matches near-field model from 0.25 m","Indoor RIS measurements at 304 GHz validate near-field gain formula","Near-field correction makes 304 GHz RIS gain model fit measurements","Phase-quantized RIS at 304 GHz fits near-field model from 0.25 m","Indoor 304 GHz RIS tests confirm near-field gain correction model"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001158,"raw_usage":{"total_tokens":4865,"prompt_tokens":1082,"completion_tokens":3783,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":698,"completion_tokens_details":{"reasoning_tokens":3686}},"tokens_in":698,"tokens_out":3783,"duration_ms":25774,"temperature":1.0,"reasoning_tokens":3686,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T18:55:07.783720+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same setup with an independent absolute calibration of the 304 GHz channel sounder (for example, a direct cable-back-to-back measurement or a known free-space path between two horn antennas) and re-measure the Tx-RIS-Rx path gain without applying any offset; if the measured curve between $d_2 = 0.25$ m and $1.15$ m deviates from formula (2) by more than the reported 2.8 dB, the validation collapses.","supporting_citations":[{"cited_title":"Channel measurements involving passive RIS at 300 GHz,","cited_arxiv_id":null,"evidence_quote":"Provides the 5.5 dB systematic correction applied to all measured path gains before comparison."},{"cited_title":"Hybrid channel model for low terahertz links in a data center,","cited_arxiv_id":null,"evidence_quote":"Describes the 304 GHz channel sounder and its measurement methodology."},{"cited_title":"A primer on near-field beamforming for arrays and reconfigurable intelligent surfaces,","cited_arxiv_id":null,"evidence_quote":"Defines the depth-of-focus distance used to interpret where far-field predictions break down."},{"cited_title":"Double directional channel measurements at 300 GHz in indoor environments enhanced by reconfigurable intelligent surfaces,","cited_arxiv_id":null,"evidence_quote":"Underlies the double-directional power angular profile measurement approach in Setup 2."}],"review_version":1}