Standard linear optics requires that the transmittance of an optic be independent of photon number: for a linear element the input–output relation a_out = t·a_in + r·b_in is an operator identity, so single-photon and bright-light transmission are equal as a theorem rather than as an approximation. We are not aware of a dedicated test of that equality with a controlled scatterer at the 10⁻⁴ level. We state a hypothesis in which the equality fails in the single-quantum sector by a fixed dimensionless coupling λ*, and specify the tabletop experiment that would exclude it. The observable is Δ = D_quantum − D_classical, where D is the half-difference of a two-arm log-ratio under an ABBA swap of a weak scatterer, measured once with counting detectors and once with photodiodes through identical geometry. Standard quantum mechanics fixes Δ = 0 exactly; the hypothesis predicts Δ = −λ*·(−p ln p), or −0.046·λ* at a 1% transmission deficit. The loss map is stated as a photon-number-dependent Kraus channel, K = Σ exp[−½λ*·I_R(n)]|n⟩⟨n| with P_R(n) = 1 − (1−p)^n, which is linear, completely positive, local and non-signalling; invariance of a remote marginal is verified explicitly on a two-mode squeezed vacuum to 3×10⁻¹⁷. An earlier thermal formulation of the same idea is excluded by published SNSPD efficiency data at 2 K by a factor of 2×10⁵ and is withdrawn. An attempt to derive the criterion distinguishing which optical elements incur the cost failed under independent adversarial review and is reported as a failure; the surviving readings are converted into five pre-registered experimental arms with their nulls stated in advance. A borrowed teaching bench at 780 nm reaches a total uncertainty of 2.9×10⁻⁴ per operating point at 2×10⁸ counts, giving 5σ sensitivity at λ* ≥ 0.031 and 95% exclusion to 0.013 at a 1% scatterer; an extended scan through the turnover at p = 36.8% reaches 0.0039 and 0.0016 respectively, subject to a costed high-loss error-budget row and seven stated pass criteria. Five signatures distinguish the effect from systematics, including a −p ln p scaling and a maximum located at p = 0.368. The cheapest arm needs no interferometer: the single-photon versus bright transmission of one 50/50 beam splitter at 10⁻³, which bounds one reading of the hypothesis below λ* = 0.003 in an afternoon’s run on a prepared bench. The protocol is pre-registered before any bench is borrowed. The result will be published in either direction; a null is reported as the best bound in this sector.
Standard linear optics requires that the transmittance of an optic be independent of photon number: for a linear element the input–output relation a_out = t·a_in + r·b_in is an operator identity, so single-photon and bright-light transmission are equal as a theorem rather than as an approximation. We are not aware of a dedicated test of that equality with a controlled scatterer at the 10⁻⁴ level. We state a hypothesis in which the equality fails in the single-quantum sector by a fixed dimensionless coupling λ*, and specify the tabletop experiment that would exclude it. The observable is Δ = D_quantum − D_classical, where D is the half-difference of a two-arm log-ratio under an ABBA swap of a weak scatterer, measured once with counting detectors and once with photodiodes through identical geometry. Standard quantum mechanics fixes Δ = 0 exactly; the hypothesis predicts Δ = −λ*·(−p ln p), or −0.046·λ* at a 1% transmission deficit. The loss map is stated as a photon-number-dependent Kraus channel, K = Σ exp[−½λ*·I_R(n)]|n⟩⟨n| with P_R(n) = 1 − (1−p)^n, which is linear, completely positive, local and non-signalling; invariance of a remote marginal is verified explicitly on a two-mode squeezed vacuum to 3×10⁻¹⁷. An earlier thermal formulation of the same idea is excluded by published SNSPD efficiency data at 2 K by a factor of 2×10⁵ and is withdrawn. An attempt to derive the criterion distinguishing which optical elements incur the cost failed under independent adversarial review and is reported as a failure; the surviving readings are converted into five pre-registered experimental arms with their nulls stated in advance. A borrowed teaching bench at 780 nm reaches a total uncertainty of 2.9×10⁻⁴ per operating point at 2×10⁸ counts, giving 5σ sensitivity at λ* ≥ 0.031 and 95% exclusion to 0.013 at a 1% scatterer; an extended scan through the turnover at p = 36.8% reaches 0.0039 and 0.0016 respectively, subject to a costed high-loss error-budget row and seven stated pass criteria. Five signatures distinguish the effect from systematics, including a −p ln p scaling and a maximum located at p = 0.368. The cheapest arm needs no interferometer: the single-photon versus bright transmission of one 50/50 beam splitter at 10⁻³, which bounds one reading of the hypothesis below λ* = 0.003 in an afternoon’s run on a prepared bench. The protocol is pre-registered before any bench is borrowed. The result will be published in either direction; a null is reported as the best bound in this sector.
Two working documents that the deposited paper (Draft 4, v1.0; DOI 10.5281/zenodo.22678044; printed title The Asymmetric-Loss Equal-Born Test: a pre-registered search for record-information-dependent optical loss) cites and undertakes to deposit alongside it. They are deposited here as they stood on September 7, 2026, the day Draft 4 was completed, each with a dated deposit note at its head; nothing beneath the note is altered. Supplement 1 — Pre-Bench Closure (September 4, 2026), with postscript layers of September 6 and 7. The arithmetic companion to the paper. Part A runs the no-signalling check on the programme's surviving statement, finds that the normalised form permits superluminal signalling and is excluded, and records the single repair — a loss channel — that leaves the bench prediction numerically unchanged and converts the hypothesis from a nonlinear modification of the Born rule into a linear one. Part B carries the consequences into the protocol: the turnover of −p ln p at p = 1/e and the extended scan that would see it, a fourth signature, the cost of the high-loss points, the total-rate observable, the absorber control with its prediction stated in advance, and the consolidated protocol amendments. Part C makes the structural repairs: locating the branching event, the J-conservation contradiction, and the cost functional, with both candidates carried and the location of the peak left to decide between them. Part D drafts the pre-registration and scores the open tasks. The September-6 layer, headed "Postscript — September 5–6", records the two independent referee passes (one blind, one adversarial), the photon-number-dependent loss map, the derivation of the record criterion put to a third referee instructed to kill it and killed on twelve counts, the pre-registered arms and the correction of a handover error; the addendum of September 7 records the closing of the high-loss row. 17 pages. Supplement 2 — The high-p row: proposal, adversarial pass, and the row as costed (September 7, 2026). The error-budget row for the extended scan (p = 10%, 36.8%, 60% and 70%) that the paper's §9 named as the single item blocking pre-registration of the extended points, kept in the three steps by which it was reached. Appendix A is the first-pass proposal exactly as written for review, errors included. Appendix B is the report of an independent adversarial referee instructed to kill it, verbatim: eight errors and six systematics the first pass had not carried. Appendix C is the row as costed after every finding was re-derived and accepted, with the seven pass criteria written into the pre-registration. Its findings: the reach at the turnover (5σ at λ* ≥ 0.0039, 95% exclusion to 0.0016) stands only under the seven criteria; the bright-phase thermal family has the sign of the predicted anomaly and is bounded by a power-dependence test stated in advance; the far point of the scan moves from 70% to 60%; and the extended scan is research-group metrology, not teaching-laboratory work. 25 pages with its deposit note. The coloured highlights in Supplement 1 are its dated layers (rose, September 6; periwinkle, September 7; sky, the deposit note of September 16, 2026), each explained at its head. The line "Not for deposit" on the cover of Supplement 2 recorded the authors' working practice on September 7 and is superseded by the deposit note above it; the paper's Draft-4 note undertakes to deposit the document. The reconciliation document the paper also cites is the programme's working record, still being expanded, and is available from the authors.
Why this measurement is worth a bench: see section 2 of the proposal, "Why anyone should want to measure it". Contact: concursion.alebt@gmail.com. Two companion documents to the deposited paper, doi:10.5281/zenodo.22678044, and its pre-registration, osf.io/3h6yf. The question: does a 50/50 beam splitter transmit a single photon in exactly the same proportion as it transmits bright light? Standard linear optics says yes, exactly. The first run compares the two to one part in a thousand, with one source; only a calibrated attenuator changes between the bright phase and the single-photon phase. ALEBT_proposal_paper_v1_oct20.pdf: the proposal written for a host laboratory (the observable, the hypothesis, the error budget, the stages, the equipment and the open problems). ALEBT_equipment_list_v2_oct26.pdf: the first-run equipment list, with answers to the questions raised by groups we have written to (existing measurements, fibre-coupled detectors, differences between detectors). Version 2, 6 November 2026. Adds ALEBT_equipment_list_v3_nov06.pdf, which opens with a new section, "The one observable in which the hypothesis and linear optics differ, and its predicted size": zero exactly in linear optics; minus lambda* times (minus p ln p) under the hypothesis, with a fixed sign and a peak at p = 1/e; for the beam splitter alone a null at one part in 1,000 bounds lambda* below 0.003. It was written because two theorists who read the material did not find that difference stated in one place. The proposal (v1, 20 October) and the equipment list v2 (26 October, with its answers on existing measurements, fibre-coupled detectors and detector differences) are carried unchanged. The paper is now at version 2.1, doi:10.5281/zenodo.24014717, which quotes the Polyakov and Migdall (2007) comparison correctly as 0.14% plus or minus 0.14% at k = 1. Version 3, 6 November 2026. The proposal itself is now version 2 (ALEBT_proposal_paper_v2_nov06.pdf), replacing ALEBT_proposal_paper_v1_oct20.pdf, which remains in the earlier versions of this record. It adds section 2.0, "The one observable in which the hypothesis and linear optics differ, and its predicted size", and corrects the Polyakov and Migdall (2007) figure to 0.14% plus or minus 0.14% at k = 1 in two places. Nothing else in the proposal is changed. The two equipment lists (v2 of 26 October and v3 of 6 November) are carried unchanged.
How can coherent light be certified when its amplitude is unknown or its preparations are correlated? This article treats these as distinct inference problems: learning an actual coherent reference with optimal source-copy cost for independent identical inputs, and certifying the photon-counting predictions of the light left unused in a preprepared correlated batch. Optimal source-copy cost with a learned reference. For a normal one-mode source prepared independently and identically, an ideal passive multiport test and heterodyne readout produce an actual amplitude α^. After displacement by this reference and ideal quantum-limited amplification of fixed gain G>1, let Pρ,α^G be the photon-number distribution and πG(k)=(G−1)k/Gk+1 the coherent benchmark. The task-specific comparison is G1−⟨α^∣ρ∣α^⟩≤TV(Pρ,α^G,πG)≤1−⟨α^∣ρ∣α^⟩. With at most N input copies, the protocol controls the unconditional event of accepting a reference whose actual total-variation error exceeds ε. For 0<ε<1/2, soundness error βs→0, and fixed coherent-source rejection tolerance βc<1, the optimal fixed-maximum source-copy count obeys log(1/βs)Nopt[−log(1−ε)]⟶1. The matching converse covers collective and adaptive protocols under the same source-copy contract. At 1% accuracy and 5% soundness and completeness errors, the finite bounds are 293≤Nopt≤572. These are source-copy bounds under ideal optical access, not an optimum for total hardware or reference energy. The actual unmeasured batch after the full record. A separate protocol allows an arbitrary preprepared correlated or entangled optical batch, including correlations with an environment. A pilot first fixes the reference; a private random subset is then tested. A multiplicative sampling inequality yields a frequentist guarantee for the average marginal photon-count error of the retained conditional state, conditioned on the complete actual measurement record. It does not replace that state by premeasurement marginals or assert the same error for its full joint distribution. The trusted detector contract requires product joint test effects and local instruments that leave retained systems and the external environment untouched. Calibration envelopes must cover the actual signal-time gates. The bounds retain infinite photon tails and charge detector efficiency, dark counts, imperfect displacement, reference drift and calibration failure. Explicit counterexamples explain why marginal detector efficiencies or product effects alone are insufficient. An ideal finite example uses 3730 prepared systems: 1491 pilot, 1339 tested and 900 retained; calibration and optical-control resources are additional. Relation to the original two-copy result. The earlier mean-energy-controlled arm-swap cancellation and full trace-distance certificate remain valid for independent, possibly nonidentical product batches. Their source and target are kept separate from the new arbitrary-batch marginal PNR guarantee. The analysis builds on established coherent-state comparison, quantum verification and sampling identities; the new connections supply the learned-reference resource law and the full-record conditional-state return. Version 1.1 is a substantial update of Certifying unused optical states with imperfect two-copy interference, first published on 31 October 2026. The integrated 24-page PDF contains three figure groups, a source-and-target comparison table, complete proofs and references. The figures illustrate proved bounds and specified theoretical protocols, not experimental measurements.
Loss can restore the vacuum entropy bound without erasing non-Gaussian structure. This paper determines when an entropy statement justifies a coherent-state approximation, when non-Gaussian structure survives optical loss, and what mode and source-request budgets are needed to use that structure. The results give quantum-optics researchers separate criteria for entropy recovery, state approximation, Gaussian concentration and conversion to a specified output. Version 1.5 — revised 10 October 2026. The 96-page PDF contains the main article, Methods, Supplementary Information and references. This revision connects the exact loss boundary to a mode-uniform near-equality certificate, finite-data validation, a conversion-cost lower bound for the same thermally repaired source, and an exact equal-entropy comparison of sources with different conversion-cost exponents. The central figures are updated accordingly. 1. Near-vacuum states can have an unbounded entropy deficit Write νm for the m-mode vacuum and define Δα,m(ρ)=Hα(Wνm)−Hα(Wρ), using natural logarithms; H1 is Shannon entropy. Positive deficit means entropy below vacuum. We construct physical Wigner-positive states satisfying, simultaneously, Em=Tr(Nmρm)→0,21∥ρm−νm∥1→0,dFm(ρm)→0,Δ1,m(ρm)→+∞. Here Fm is the trace-norm-closed convex hull of Gaussian states and dFm uses half the trace norm. Thus proximity to the entire Gaussian model does not control Wigner entropy uniformly in mode number. Independent weak modes amplify a one-mode entropy effect while their total excitation vanishes. For the same sequence, half loss instead gives −Em≤Δ1,m(L1/2⊗mρm)≤0, so the entropy difference from vacuum tends to zero. 2. The exact pure-loss recovery boundary For Wigner-positive inputs, intensity transmissivity η and any fixed positive total input-energy budget E, the sharp threshold is ηc(α)={1/2,α/2,0<α≤1,1<α<2. With Pm,E={ρ:Wρ≥0,Tr(Nmρ)≤E}, supm≥1supρ∈Pm,EΔα,m(Lη⊗mρ)={0,+∞,η≤ηc(α),η>ηc(α). The recovery statement includes the boundary itself and is independent of mode number. Half loss also restores vacuum majorization for arbitrary density operators. For α≥2, established Wigner-norm bounds give the vacuum entropy bound at every transmissivity for Wigner-positive outputs. These statements distinguish a universal guarantee from the earlier sign crossing that a particular source may exhibit. 3. Near equality gives a certificate for the entire input For an arbitrary input ρ, define the half-loss output Shannon excess s(ρ)=h(WL1/2⊗mρ)−m(1+lnπ)≥0. The half-loss identity identifies this quantity with the input's total Wehrl entropy excess. Applying mode-uniform Wehrl stability yields infa∥ρ−∣a⟩⟨a∣∥12≤ln264s(ρ). The bound controls the full input state, with no energy or photon-support restriction. One coherent reference then controls every subsequent quantum channel and every unconditional measurement event. In particular, an event-probability error at most ϵ is guaranteed by s≤(ln2)ϵ2/16. Conversely, a surviving output Gaussian-hull gap enforces s(ρ)≥16ln2dFm(L1/2⊗mρ)2. Recovery of the inequality and proximity to equality therefore support different physical inferences: the former permits residual non-Gaussian structure, while the latter certifies a coherent-state approximation. 4. Surviving structure has mode and conversion costs A thermally repaired vacuum–three-photon source with per-mode occupation c/m has total input energy tending to 3c. Along the recovery boundary, its attenuated state obeys liminfm→∞dFm(ρm,ηc)≥cηc3exp{−c[1−(1−ηc)3]}>0. At half transmission this lower bound is g(c)=(c/8)e−7c/8. Starting from one complete block, Gaussian processing requires Θ(m) output modes to retain any fixed positive gap below this bound at success probability at least a fixed p0>0. Gaussian ancillary states, adaptive Gaussian measurements, feedback, discarding and postselection are included. A distinct task requests fresh independent blocks of this same half-loss source to produce the one-mode target τ∗=(18∣0⟩⟨0∣+6∣1⟩⟨1∣+∣2⟩⟨2∣)/25 within half trace distance 10−6. Let Cm be the infimum of expected complete-block requests divided by success probability, including all failed requests. For fixed 0<c≤1, every achieving Gaussian protocol obeys the asymptotic lower bound logCm≥1281(m/c)1/3−logm−O(1). This is a source-request lower bound; the optimal achievable cost for this source remains undetermined. The proof evaluates a curvature-controlled input certificate for the actual thermally repaired state and applies the Gaussian conversion theorem of the companion paper cited below. 5. Exactly equal entropy does not fix conversion cost For two separate two-mode quartic source families, their mixture probabilities can be adjusted so that their half-loss Shannon entropies are exactly equal. Under a common target, tolerance and failure-inclusive accounting, the inherited matching conversion laws become h(WωA)=h(WωB)=2(1+lnπ)+s,logCA(s)=Θ(s−1/3),logCB(s)=Θ(s−1/2). The target tolerance is held fixed in these asymptotics. The delivered energies after entropy matching satisfy EA,EB∼s/2. At exactly equal energy, the two excesses instead share si(E)=2E−(35/8)E2+O(E3) and differ by sA(E)−sB(E)=15E3+O(E4). Entropy alone therefore cannot select the cheaper source for a fixed output task. 6. Finite data and finite mode budgets Under a declared displaced-input photon-tail bound Tr[N1N>Kσa]≤τK, trusted half loss and all-mode on/off detection with efficiency at least ξmin, the click probability q gives s(σ)≤AKq+τK,AK=1−(1−ξmin/2)KK. A one-sided binomial bound turns this into an entropy confidence certificate. An explicit undisplaced four-photon-cutoff example uses K=4, τK=0, ξ=0.8 and N=3,517,336 independent validation blocks, accepting at most 41 clicks, to certify unconditional event error at most 0.05 with false-acceptance probability at most 0.001 and missed acceptance at most 0.001 on the stated low-click alternative class. All vacuum results count; training, calibration and source heralding are separate costs. When coherent-state approximation alone is the task, the same click measurement supplies a stronger direct fidelity test. The paper also gives all-input finite-budget exclusions above the loss-recovery line. For example, at E=1, α=1.9, η=0.9505 and m=105, the certified bound Δα<0.06942 rules out a target deficit of 0.1 nat. A separate finite single-mode preparation model predicts positive Shannon deficit near η=0.97, negative deficit near η=0.85, and a Gaussian-hull witness above 4.4×10−3 at half transmission, with explicit preparation, measurement-error and raw-trial budgets in the PDF. 7. Dimensional scales, a strict entropy floor and ORCS Near order two, put α=2−δ and L=ln(1/δ). For the maximum deficit Dm,E at fixed total energy, Dm,E(2−δ)=Θm,E(δ/L)for fixed m,Th(δ,E)=ΘE,h(L/δ), where Th is the first mode number allowing a fixed positive deficit h. Beyond a sufficiently large fixed multiple of L/δ, Dm,E(2−δ)=ΘE(mδ). The manuscript separates proved sufficient fronts from necessary exit bounds; the sharp first-crossing coefficient remains open. At every fixed mode number, an energy-unrestricted strict Shannon floor also holds: h(Wρ)≥mln(2π)+εm with εm>0 for all Wigner-positive states. The ordering-resolved coherence spectrum (ORCS) supplies an exact single-mode loss coordinate. With Cr,k denoting the ordering-weighted norm of photon-number-difference block k, Cr,k(Lηρ)=CTη(r),k(ρ),Tη(r)=1−η1−r. For vacuum–n-photon coherence and 1<α<2, C1−α/2(Lη∣n⟩⟨0∣)=(α2η)n. This formula exposes the high-order coherence mechanism at the loss threshold; the universal recovery result is proved through the full nonlinear Wigner-norm inequality. Related papers and history The counterexample paper Wigner entropy below vacuum: physical counterexamples and stability limits expands arXiv:2609.13312v1, submitted on 10 September 2026. This stability paper develops the dimensional, loss-dependent and operational consequences. Its state-approximation and conversion interfaces use the companion results From phase-space entropy to coherent-state error: mode-uniform Rényi–Wehrl stability and Phase records control the conversion cost of non-Gaussian light. The mode-retention comparison connects to Limits to Gaussian concentration of non-Gaussian light.
Detecting non-Gaussian light establishes that a resource is present. Its structure and preparation record determine the cost of turning it into a specified output. This paper gives matching conversion-cost laws for weak optical sources and a general input certificate that can exclude every allowed Gaussian receiver before a circuit is designed. Sources with identical energy, covariance and total photon statistics can have different cost exponents; losing a classical phase bit can make a fixed-source conversion impossible. Source, target and accounting. The Wigner-positive two-mode sources are obtained by half loss on each mode: ρt,e,q=L1/2⊗2[1+t2∣00⟩⟨00∣+t2{(1−q)∣p+⟩⟨p+∣+q∣p−⟩⟨p−∣}],∣p±⟩=1+e2∣22⟩±e(∣40⟩+∣04⟩)/2,0≤e≤41,0≤q≤21,E=1+t22t2. Here e selects the source, q is an unflagged phase-record error probability, and E is delivered mean photon number. Conversion targets τ∗=(18∣0⟩⟨0∣+6∣1⟩⟨1∣+∣2⟩⟨2∣)/25,21∥σout−τ∗∥1≤δ,0<δ<10−5. Joint processing of independent supplied copies, Gaussian ancillas, squeezing, memory, Gaussian measurements, feedback, discarding and postselection are allowed. Cost is C=EK/ps: every source request counts, including vacuum requests and failed attempts. The task uses ideal Gaussian operations and charges supplied copies. 1. A source-selection law controlled by record accuracy. Optimizing both the source parameter e and the allowed Gaussian protocol gives, uniformly in q∈[0,1/2] as E→0, logCbest(E,q)=Θδ(max{E−1/3,q/E}). Errors of order E1/3 or smaller preserve the fastest logarithmic cost. Above that scale, record quality controls the optimum. Equivalently, for q=q0Ea, fixed 0<q0≤1/2 and a≥0, logCbest=Θδ(E−β),β=max{1/3,(1−a)/2}. At fixed nonzero error, choosing the phase-insensitive source keeps conversion possible at logarithmic cost of order E−1/2. 2. The same elementary diagnostics, exponentially different costs. The normalized excitation polynomials pA(z)=48z14−z24,pB(z)=8(z12−z22)2 give calibrated half-loss sources with identical energy, covariance and total photon-number statistics, but for the same target and tolerance, logCA(E)=Θδ(E−1/3),logCB(E)=Θδ(E−1/2). The distinction is the local structure of their excitation-polynomial zeros. Full Gaussian-protocol lower bounds and matching heterodyne constructions establish both orders. 3. Certification survives errors that obstruct conversion. In the symmetric quartic family, the full first-mode photon distribution is independent of q. In particular, Px(1)=16(1+e2)E(4+e2)≥E/5,Px(N≥5)=0. Every state in the full one-mode Gaussian convex hull instead satisfies P(N≥5)≥4608000P(1)6. Photon counting therefore supplies a sufficient O(E−6) request budget at fixed false-positive and missed-detection probabilities against independent repetitions of any one unknown Gaussian-hull state. Yet, in the joint limit t→0, e/t→∞, finite-cost conversion has the threshold qc(t,e;δ)=Θδ((E/e2)1/3). For fixed e>0, any fixed positive unflagged error thus eventually obstructs this output. Flagged erasures behave differently: independently retaining a correct bit with probability ar>0 adds at most a factor 1/ar to a protocol that discards erasures. 4. A transferable exclusion certificate and an explicit receiver. A verified positive phase-space minorant with a physical pure Gaussian curvature reference supplies a source budget bκ. For a requested one-mode output whose trace-distance gap from the Gaussian convex hull is at least g, every allowed Gaussian protocol obeys C≥bκg2−κ/(2−κ),2−κκ<g2. The manuscript constructs the certificate and explicit incomplete-gamma budgets for the quartic sources. It also gives a centred-root heterodyne receiver. In its specified joint limit, with scaled error λ=deq(e/t)2/3, its optimized limiting one-photon population is supw=2[1+43−3/4λ3/4]1. The paper supplies de, the receiver action, and a photon-statistical output witness. These receiver formulas connect the desired output to tolerable record error; the general certificate supplies the protocol-independent cost bound. Together they let source designers compare preparation structure, record quality and total request budget for the same optical task.