REVIEW 8 minor
Maximum brightness theorem for waves
T0 review · 0 major / 8 minor · reviewed 2026-08-16 · deepseek-v4-flash
Pith's one-line read For any passive linear wave system, the ordered component powers of a partially coherent field can only stay the same or decrease.
desk verdict A clean, genuinely stronger componentwise brightness theorem for passive wave systems; the proof holds up, with only a minor heuristic gap in the finite-dimensional truncation. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The coherency matrix $\rho = \sum_i P_i |\chi_i\rangle\langle\chi_i|$ encodes the partially coherent field, with its eigenvalues being the separable modal powers. The proof machinery is the singular-value decomposition of the system matrix, $M = VDU$; passivity enters as $s_i^2 \le 1$ for the singular values, so $D^\dagger D \le I$. The unitary factors preserve the coherency-matrix eigenvalues, the diagonal factor supplies a Hermitian-matrix inequality between the transformed input and output matrices, and the monotonicity theorem for ordered eigenvalues converts that matrix inequality into the componentwise eigenvalue bounds of Eq. (1).
What would settle it
Measure the input and output coherency matrices, in the same ordered orthogonal basis, of a passive device that strongly mixes many modes; if any output eigenvalue exceeds the corresponding input eigenvalue, the theorem is false. Alternatively, find a passive device whose singular-value decomposition in standard orthonormal input and output mode sets has a squared singular value greater than 1, which would show the passivity-as-contraction assumption is normalization-dependent.
Extended reading notes
Core claim
The central claim is Eq. (1): if a passive linear wave system is represented by a matrix $M$, and the input and output coherency matrices $\rho_S$ and $\rho_R = M\rho_S M^\dagger$ have eigenvalues $\mu_{Si}$ and $\mu_{Ri}$ arranged in non-increasing order, then $\mu_{Ri} \le \mu_{Si}$ for every $i$. The paper states this as the Maximum Brightness Theorem: in passing through a passive optical system, the power in the $i$-th most powerful mutually incoherent, mutually orthogonal component of the field cannot increase. The proof uses the singular-value decomposition $M = VDU$ with $D^\dagger D \le I$, compares the intermediate coherency matrices as Hermitian matrices under the partial order in which $A\preceq B$ means $B-A$ is positive semidefinite, and applies the standard monotonicity theorem for ordered eigenvalues to conclude that every ordered eigenvalue can only stay the same or decrease.
Load-bearing premise
The theorem stands or falls on the assumption that the field can be represented in a finite set of modes and that every channel of the passive system has gain at most one; if either fails, the matrix comparison that drives the proof does not apply.
Editorial extensions
If this is right
- No passive optical system can concentrate partially coherent radiation into a single output mode beyond the power of the strongest orthogonal incoherent input component.
- For a system with known squared singular values $s_i^2$, the largest possible separable output power is bounded by $s_1^2\mu_{S1}$, and the largest possible total output power by $\sum_i s_i^2\mu_{Si}$, with optimal coupling obtained by aligning input coherent modes to the system's communication modes.
- The threshold rank, defined as the number of coherent-mode powers above a threshold $\epsilon$, cannot increase in a passive system and is conserved in a lossless system.
- The theorem forbids any passive apparatus that redistributes the sum of squared coupling strengths to create more moderately coupled orthogonal channels, because such a redistribution would raise some ordered eigenvalue.
- The bound applies to scattering-matrix descriptions and non-reciprocal systems, so it constrains radio-frequency energy harvesting, antenna systems, and nanophotonic concentration as well as conventional imaging optics.
Reading between the lines
- An experimentalist could test the bound with a programmable interferometer mesh by measuring both coherency matrices for many random passive settings; an ordered violation would signal a mode-truncation artifact rather than a failure of the mathematical claim.
- Because the proof relies only on passivity and linearity, an analogous componentwise bound should hold for time-dependent linear passive media analyzed frequency band by frequency band, although the paper does not develop that extension.
- Read as an unnormalized density matrix, the coherency matrix makes the theorem a statement about ordered photon-number probabilities in orthogonal modes under passive linear optics, connecting it to majorization-based reasoning in quantum information that the paper only touches on implicitly.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This paper introduces a "Maximum Brightness Theorem" for passive linear wave systems. For an input coherency matrix ρ_S and a passive transformation M, the output coherency matrix is ρ_R = Mρ_SM†; the theorem states that the eigenvalues of ρ_R, ordered non-increasingly, are bounded componentwise by the eigenvalues of ρ_S. The proof combines the singular-value decomposition M=VDU, unitary invariance of the spectrum, the Loewner order, and Weyl's monotonicity theorem. The paper also derives consequences: a threshold rank bound, a maximum total brightness bound, and a thought experiment showing that the number of moderately coupled channels cannot be increased. The authors provide supplementary proofs of the auxiliary matrix facts and discuss connections with majorization and classical brightness/étendue.
Significance. The theorem is a very clean and general statement that strengthens earlier results by Zhang, Hsu, and Miller (ref. 1) and the majorization bound (ref. 2). It applies to arbitrary passive linear systems, including lossy and non-reciprocal ones, and to partially coherent fields. The proof is elementary and self-contained, with no free parameters and no circularity. The physical consequences—bounds on single-mode concentration of partially coherent light, total transmitted power, and threshold rank—are useful and clearly explained. The paper is honest about relying on communication-mode theory for the finite basis justification. If the infinite-dimensional scope is clarified, this should become a standard reference.
minor comments (8)
- [Abstract and Methods] The theorem is stated for "arbitrary passive optical or wave systems" (abstract, Eq. (2)), while the proof in the Methods assumes an n×n matrix M (Eqs. (14)–(16)) and uses the finiteness of n for Weyl's monotonicity theorem. Please add a sentence clarifying that the proof is for finite mode bases and that the infinite-dimensional case follows by a standard limiting argument (or, alternatively, explicitly restrict the theorem to finite-dimensional mode truncations).
- [Methods, Eq. (25)] If the singular values s_i are allowed to be complex (as stated in the paragraph before Eq. (21)), the diagonal entries of D†D should be |s_i|^2, not s_i^2, and the passivity condition in Eq. (16) should be written as |s_i|^2 ≤ 1.
- [Supplementary text S3] The proof of the cyclic property establishes equality only for non-zero eigenvalues. Since the Methods later compares complete ordered eigenvalue lists, please cite the standard result that square matrices AB and BA have identical characteristic polynomials, or extend the proof to include zero eigenvalues.
- [Methods, Eqs. (22)–(26)] Please define I_n explicitly as the n×n identity matrix and note that D†D is diagonal with entries |s_i|^2, so that the Loewner comparison D†D ⪯ I_n is immediate from |s_i| ≤ 1.
- [Main text, after Eq. (4)] There is a duplicate phrase: "for any χ in the space of interest in the space of interest" should be "for any χ in the space of interest."
- [Supplementary text S5] The displayed line "So, SwRμμ " appears garbled and should be typeset as a proper statement of weak majorization.
- [Eq. (3)] Please state the normalization convention for the basis functions that makes the eigenvalues of the coherency matrix equal to optical powers (e.g., power-normalized modes). This convention is implicitly assumed when Eq. (16) is called a passivity condition.
- [Discussion, S7] The theorem is presented for a square n×n matrix M. If input and output spaces have different dimensions, please clarify whether the shorter eigenvalue list is padded with zeros or whether equal dimensions are assumed after truncation.
Circularity Check
No significant circularity: the eigenvalue bound is derived from SVD, Loewner order, and Weyl monotonicity, with no fitted parameters and no assumption equivalent to the conclusion.
full rationale
Walking the derivation chain: the paper defines input/output coherency matrices by the standard propagation law ρ_R = M ρ_S M† (Eq. 6, proved in S1). It then decomposes M by SVD (Eq. 14), uses passivity only to assert s_i^2 ≤ 1 (Eq. 16), and proves D†D ≤ I (Eq. 26). The Loewner comparison R' ≤ ρ̃_S (Eqs. 29-34) follows by elementary algebra from ρ̃_S^{1/2} D†D ρ̃_S^{1/2} ≤ ρ̃_S, using the cyclic property and the square root of a positive semidefinite matrix. Weyl's monotonicity theorem (S4, proved from the Courant-Fischer minimax theorem) then converts the Loewner order into the ordered eigenvalue inequalities (Eq. 35), i.e., μ_Ri ≤ μ_Si. None of these steps assumes the target inequality; the only physical input is passivity/no-gain, stated as s_i^2 ≤ 1, which is an explicit premise rather than a consequence of Eq. (1). No parameter is fitted and no 'prediction' is a renamed fit. The self-citations (refs 5, 10, 19) are used only to justify the practical finite-n truncation and the existence of communication modes; even if those citations were set aside, the finite-dimensional inequality holds for any n×n contraction, and the same Weyl argument extends to infinite-dimensional trace-class coherency operators under the same contraction condition. Thus the central claim has independent mathematical content and is not circular.
Assumptions & free parameters
assumptions (7)
- standard math Singular value decomposition of any linear matrix M
- standard math Weyl monotonicity theorem
- standard math Cyclic property of matrix products
- standard math Square root of a Hermitian positive semi-definite matrix
- domain assumption Passivity implies all singular values satisfy s_i^2 <= 1
- domain assumption Finite-dimensional truncation of the wave field to n modes
- domain assumption Quasi-monochromatic or narrowband decomposition
Cite this review
Pith. "Pith review of Maximum brightness theorem for waves." pith.science (2026). https://pith.science/paper/U4N7G3O3
@misc{pith2026260812578,
author = {Pith},
title = {Pith review of: Maximum brightness theorem for waves},
year = {2026},
howpublished = {\url{https://pith.science/paper/U4N7G3O3}},
note = {Machine review of arXiv:2608.12578}
}
read the original abstract
We introduce a universal bound on the separable powers of a wave field after passing through arbitrary passive optical or wave systems. Any wave field can be expressed as a combination of mutually incoherent and mutually orthogonal components, each with some power or "brightness". We prove that, in passing through a lossless or lossy optical system, writing the components in order of power, the power in each such component at the output cannot exceed the power in each such component at the input, even though each resulting output may be an arbitrary mixture of the inputs. This result encompasses previous brightness theorems, has several immediate consequences, and gives a simple limit to the concentration of light, radio-frequency or other waves into single-mode outputs.
Figures
Reviewed August 16, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.