{"id":"7014cf76-8270-45cf-a8f7-c22cf8ab3bdc","arxiv_id":"2607.15138","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A time-domain point-and-shoot method translates spherical-harmonic acoustic data between non-intersecting spheres using Kirchhoff-Helmholtz ring integrals, with six-digit accuracy in a parameter window.","lead":"Acoustic fields are shifted from one sphere to another using spherical-harmonic data and a Kirchhoff-Helmholtz surface integral, with rotations before and after an axial translation. The method targets time-domain fast multipole calculations and reports six-digit accuracy and about quadratic scaling for practical parameter ranges.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The abstract's O(N^2) scaling claim is internally inconsistent with the paper's own algorithm and timing data.","rationale":"The Reader's weakest_assumption focused on the approximate rotation operator (Eq. 12) and its missing error bound. That is a legitimate correctness risk, but the empirical tests already demonstrate six-digit accuracy for the tested source, where high-degree coefficients decay rapidly; the rotation concern is therefore a generality gap rather than an observed failure. The complexity inconsistency is more concrete and central: the abstract's headline performance claim is directly contradicted by the paper's own Section 2.7 and Figure 12. The axial translation loop structure is O(N^3) under the natural scaling of quadrature rings, and rotation is O(N^4); there is no algorithmic device in the paper that reduces these to O(N^2). If the timing data show N^2.1, it must be because the ring count is fixed or the range of N is too small, meaning the claimed scaling is not a true scaling law. This warrants at least a conditional acceptance with a required revision of the abstract and a proper complexity analysis. The paper's contribution remains credible and the six-digit convergence for suitable parameters is supported by the presented tests, so a rejection is not justified. The Reader mentioned the complexity issue in the rationale but did not make it the primary weakness, hence 'partial' agreement.","tokens_in":9340,"tokens_out":10121,"duration_ms":109131,"concrete_test":"Instrument the code to record the number of source and target rings used for each N in the Figure 8/12 tests. Then rerun the timing experiment for N = 32, 40, 48, 64 with n_t = 512, both keeping the ring counts fixed and scaling them with N. Fit log-log slopes for translation, rotation, and total time. If ring counts grow linearly with N and the translation slope moves from ~2 toward ~3 (or total toward ~3.5), the abstract's O(N^2) claim is false. At minimum, inspect Algorithm 4's loop bounds and count the ring-pair–mode products analytically.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central performance claim — 'computational effort scales approximately as N^2' — is contradicted by the paper's own algorithm and measurements. Section 2.7 states that rotation matrices are N(N+1) × N(N+1), implying O(N^4) per time step, and Figure 12 fits rotation time as N^3.5. The axial translation in Algorithm 4 uses three nested loops over source rings, target rings, and azimuthal modes, so it is O(R_s R_t N). If ring counts scale with N (as required for quadrature accuracy), this is O(N^3), not O(N^2). Section 2.7's argument that 'generation of azimuthal modes' and 'transfer of N source modes' are each O(N^2) silently omits the spatial quadrature over rings. The measured N^2.1 translation slope therefore cannot be an asymptotic scaling law unless the number of rings is held fixed, which would under-resolve higher N. This is an internal inconsistency in a load-bearing assertion of the abstract, independent of the rotation-accuracy issue identified by the Reader.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a method to translate transient acoustic field data between two non-intersecting spheres with arbitrary relative orientation. The field on each sphere is represented as a truncated spherical harmonic expansion. The translation uses a 'point-and-shoot' strategy: rotate the source expansion into a frame aligned with the source-to-target axis, perform an axial translation using precomputed Kirchhoff–Helmholtz ring impulse responses, and rotate the resulting target expansion back. Numerical tests compare the translated coefficients against direct evaluation from a point source and report close to six-digit accuracy for a range of parameters, except for the very closest interaction. The paper claims computational effort scales approximately as N^2, where N is the spherical harmonic order.","tokens_in":9653,"tokens_out":3211,"duration_ms":34365,"significance":"If the method is both accurate and truly O(N^2) per time step, it would provide a practical translation operator for time-domain fast multipole methods, filling a known gap. The paper has several genuine strengths: the Kirchhoff–Helmholtz derivation is standard, the error is measured against a direct analytic point-source evaluation with no fitted constants, and the method is presented in algorithmic detail with precomputation steps. The author's honest reporting of the failure for the closest interaction is also a positive feature. However, the complexity claim is contradicted by the paper's own analysis and timing data, and the rotation step lacks a rigorous accuracy guarantee, so the central practical claim needs significant revision.","major_comments":[{"comment":"The central claim 'computational effort scales approximately as N^2' is inconsistent with the paper's own complexity analysis and measurements. Section 2.7 states that rotation matrices are of size N(N+1) × N(N+1), which implies O(N^4) work per matrix-vector multiplication, and Fig. 12 fits the rotation time as N^3.5. Even if the constant is small, the asymptotic scaling is not N^2. The abstract and conclusions should be corrected to state the actual scaling, or the rotation must be replaced by an algorithm with provably lower complexity.","section":"Abstract; Sec. 2.7; Fig. 12"},{"comment":"The claimed O(N^2) for the axial translation is not supported by the algorithm as written. Algorithm 4 contains three nested loops over source rings i, target rings j, and azimuthal modes m. If the number of rings on each sphere grows with N, as required for accurate quadrature, this loop structure is O(R_s R_t N) = O(N^3), not O(N^2). The text in Sec. 2.7 counts only the 'generation of azimuthal modes' and 'transfer of N source modes' but omits the double sum over rings. The measured N^2.1 slope may reflect a fixed or slowly growing ring count over the tested range, but it cannot be taken as an asymptotic scaling law.","section":"Sec. 2.4; Algorithm 4"},{"comment":"The rotation method R = A Y_R, following Lessig et al., is exact only if the quadrature rule used to build A remains exact after the arbitrary rotation of the evaluation nodes. For Lebedev-type quadratures, exactness relies on the specific node set and weights; rotating the nodes does not in general preserve exactness. Since every translation applies both a forward and an inverse rotation, any rotation error propagates into the six-digit convergence claim. The paper offers no error bound or separate convergence test for the rotation. I ask that the authors quantify the rotation error (e.g., by rotating a known expansion through a set of angles and measuring the coefficient error) and either show it is negligible in the reported parameter window or replace the rotation with an exact Wigner-D based method.","section":"Sec. 2.5; Eq. (12)"}],"minor_comments":[{"comment":"The notation in Eq. (15) uses Q(1)_m and N(1)_m for source ring modes but the target coefficients are denoted Q(2)_m and N(2)_m. This is understandable but should be made explicit, especially in the Fourier coefficient recovery step that follows.","section":"Sec. 2.4, Eq. (15)"},{"comment":"Typo: 'Algorith 1' should be 'Algorithm 1'. Also, the ordering of loops over i and j is not described; clarifying whether the impulse responses are stored as a flat array would improve reproducibility.","section":"Algorithm 4"},{"comment":"The number N_φ of azimuthal quadrature points is never specified or analyzed. Since the ring impulse responses are computed with this discrete sum, its accuracy is part of the overall convergence. Please state the values used in the tests and provide guidance on how N_φ should scale with N.","section":"Sec. 2.3, Eq. (14)"},{"comment":"The figure reports computation time for rotation and translation separately, but the translation time appears to include only the axial step, not the two rotations. Since the abstract claims N^2 for the whole operation, the total time (rotations plus translation) should be plotted and fitted separately, or the wording clarified.","section":"Sec. 3, Fig. 12"},{"comment":"Data availability: 'Code implementing the method... is available upon request' is a weaker statement than providing a permanent repository. Given that the numerical claims are central, a versioned code release would substantially strengthen the paper.","section":"Sec. 5"}],"recommendation":"major_revision","confidential_remarks":"The core translation scheme appears technically plausible and the numerical experiments support six-digit accuracy in a parameter window. The main obstacle is the unsupported and internally contradicted complexity claim, which is a selling point of the paper. If the authors can either prove the O(N^2) scaling under explicit assumptions on quadrature sizes or honestly re-state the complexity as O(N^3)-to-O(N^4) with empirical constants, and if the rotation accuracy is quantified, the paper would be publishable. The current version, however, overstates its central contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read it. The one thing to know: the core translation idea is real and the numerical section is mostly honest, but the abstract oversells the scaling.\n\nWhat's new: the paper translates time-domain spherical-harmonic expansions between arbitrarily oriented, non-intersecting spheres directly, without Laplace transforms or concentric geometry. The Kirchhoff-Helmholtz ring-to-ring impulse responses are precomputed; the usual point-and-shoot rotations are then matrix-vector products per time step. Error tests cover local, upward, and downward shifts, and show roughly six-digit errors inside a parameter window. The paper is also honest that the closest local interaction fails and would have to be handled by direct evaluation. That is a useful, credible numerical primitive.\n\nWhat worries me: the abstract's 'approximately N^2' claim does not survive contact with the paper. Section 2.7 says the rotation matrices are N(N+1) by N(N+1), which is O(N^4), and Figure 12 fits rotation time as N^3.5. Figure 12 fits axial translation as N^2.1, but Algorithm 4 is nested over source rings, target rings, and azimuthal modes. If the number of rings grows with N, that loop is closer to O(N^3). The measured slope describes the tested range, not an asymptotic law. The rotation itself is approximate—evaluation at rotated quadrature nodes—and no error bound is supplied, even though every translation uses two rotations. Less serious: only one source configuration is tested, the printed axial expressions are worth checking for a surface-area normalization factor, and code is only 'available upon request.'\n\nBottom line: the method is a legitimate contribution for time-domain FMM/BEM implementers and deserves a serious referee. The referee should ask for a corrected complexity statement, some rotation error analysis or experiments, and a second source distribution. I would cite it if I worked in this area, and I would bring it to reading group mainly to discuss the scaling gap.","headline":"A genuinely direct time-domain SH translation operator with honest error tests and a plausible core, but the paper's N^2 complexity claim is internally inconsistent with its own algorithm and scaling data.","tokens_in":10085,"tokens_out":2872,"would_cite":true,"duration_ms":31353,"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":"A point-and-shoot translation method uses the Kirchhoff–Helmholtz integral to shift spherical harmonic acoustic expansions between spheres, converging to six digits and scaling as N^2.","keywords":["acoustic field translation","spherical harmonics","fast multipole method","Kirchhoff–Helmholtz integral","time-domain acoustics","point-and-shoot","transient waves","Lebedev quadrature"],"falsifier":"Take a source sphere with expansion order N=40, translate by the (2,1,0) shift with a rotation angle that does not align with quadrature symmetries, and compare the translated expansion against direct evaluation from the source for a point source. If the coefficient error does not reach about 1e-6 for sphere radius a=1.2*sqrt(3)d while increasing N to 48, the rotation or translation assumption fails. Alternatively, rotate a known band-limited spherical harmonic expansion by an arbitrary angle using R=AY_R and check whether the coefficients are preserved to machine precision as the number of qu","tokens_in":9256,"feed_emoji":"🎯","tokens_out":3861,"duration_ms":38624,"temperature":0.7,"pith_summary":"The paper aims to establish a practical translation operator for time-domain fast multipole methods: given a spherical harmonic expansion of transient acoustic data on one sphere, it computes the expansion on another, arbitrarily oriented sphere. The method splits the shift into a rotation, an axial translation via precomputed ring impulse responses, and a rotation back. The author shows numerically that the error converges to six digits for suitable sphere radii and that per-time-step cost scales approximately as N^2, where N is the spherical harmonic order. If correct, this removes a known bottleneck in transient FMM and makes time-domain translations comparable in cost to frequency-domain ones.","feed_headline":"Acoustic field shifts converge to six digits in N² time","feed_subtitle":"A point-and-shoot scheme using the Kirchhoff–Helmholtz integral makes time-domain fast multipole translations practical.","key_machinery":"The key object is the set of precomputed weighted impulse responses I^(m)_i (r1,r2,z) for each source-target ring pair, derived from the Kirchhoff–Helmholtz integral and advanced-time Lagrange weights. Along with the rotation matrices R = A Y_R (evaluating the expansion at rotated nodes), these operators turn the shift into a sequence of matrix-vector products: rotate, apply axial translation ring by ring, rotate back. The axial translation is O(N^2) because the azimuthal mode order is preserved, and only N source modes are transferred to N target modes per ring.","core_discovery":"The central claim is that the Kirchhoff–Helmholtz integral, applied to spherical harmonic surface data, gives a complete shift operator for transient fields: after rotating the source expansion so the translation axis is the z-axis, the integral factorizes over coaxial rings, each modal interaction being a precomputable impulse response. Applying the advanced-time Lagrange interpolation, these responses update the target modal coefficients at each time step. Two rotations, implemented by resampling at rotated quadrature nodes, handle arbitrary orientations. The method is shown to converge to six digits for local, upward, and downward shifts when sphere radii are chosen away from the closest","pith_inferences":["If the rotation step error is bounded as N grows, the same point-and-shoot structure could extend to other wave equations, such as vector or elastic waves, by replacing the scalar Kirchhoff–Helmholtz kernels with their tensorial analogues.","The observed N^2 scaling suggests the axial translation is essentially a convolution in azimuth; one might replace the ring loop with a fast transform to push the exponent lower for very high N.","The failure of the (2,0,0) case hints that accuracy is controlled by the ratio of sphere radius to box size; a user could choose sphere radii adaptively per interaction rather than globally.","One could test the rotation component in isolation by rotating a known band-limited function and measuring coefficient error as a function of quadrature order, which would separate rotation error from translation error."],"forward_implications":["Time-domain FMM shift operations become practical at expansion order up to about 48, with per-step cost scaling as N^2.1 and a small leading constant.","Six-digit accuracy holds for local, upward, and downward shifts in a parameter window of sphere radius roughly 1.15 to 1.25 times the minimum enclosing radius.","The very closest local interaction, the (2,0,0) shift, is not computed accurately and would need direct source evaluation; the method is limited to non-overlapping spheres.","The precomputed impulse responses are reusable across all time steps, making the method well suited to long transient simulations."],"fun_headline_variants":["Point-and-shoot acoustic translation hits six-digit precision","Acoustic field translation: point-and-shoot to six digits","Time-domain multipole shifts: N² cost, six-digit accuracy","Shift acoustic fields fast: six-digit convergence, N² scaling","Kirchhoff–Helmholtz acoustic shifts: six digits, N² effort"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The rotation of spherical harmonic expansions is done by resampling the function at rotated quadrature nodes with the same integration weights, which is exact only if the quadrature rule stays exact after an arbitrary rotation—something the paper does not prove or bound.","fun_headline_variants_meta":{"raw":{"variants":["Point-and-shoot acoustic translation hits six-digit precision","Acoustic field translation: point-and-shoot to six digits","Time-domain multipole shifts: N² cost, six-digit accuracy","Shift acoustic fields fast: six-digit convergence, N² scaling","Kirchhoff–Helmholtz acoustic shifts: six digits, N² effort"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000765,"raw_usage":{"total_tokens":3162,"prompt_tokens":612,"completion_tokens":2550,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":356,"completion_tokens_details":{"reasoning_tokens":2461}},"tokens_in":356,"tokens_out":2550,"duration_ms":17701,"temperature":1.0,"reasoning_tokens":2461,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T00:02:46.504955+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a source sphere with expansion order N=40, translate by the (2,1,0) shift with a rotation angle that does not align with quadrature symmetries, and compare the translated expansion against direct evaluation from the source for a point source. If the coefficient error does not reach about 1e-6 for sphere radius a=1.2*sqrt(3)d while increasing N to 48, the rotation or translation assumption fails. Alternatively, rotate a known band-limited spherical harmonic expansion by an arbitrary angle using R=AY_R and check whether the coefficients are preserved to machine precision as the number of qu","supporting_citations":[],"review_version":1}