arxiv: 2603.10813 · v2 · submitted 2026-03-11 · 🧮 math.SP

Recognition: 2 theorem links

· Lean Theorem

Spectral deviation of concentration operators on reproducing kernel Hilbert spaces

Felipe Marceca , Jos\'e Luis Romero , Michael Speckbacher , Lisa Valentini

Authors on Pith no claims yet

Pith reviewed 2026-05-15 13:32 UTC · model grok-4.3

classification 🧮 math.SP

keywords concentration operatorsreproducing kernel Hilbert spacesspectral deviationplunge regionGabor multipliersshort-time Fourier transformeigenvalue distributiondiscretization

0 comments

The pith

Concentration operators on reproducing kernel Hilbert spaces have eigenvalue plunge regions whose size transfers uniformly from continuous to discrete settings on fine grids.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that the eigenvalue profile of concentration operators on reproducing kernel Hilbert spaces can be estimated in a manner that applies simultaneously to continuous and discrete cases. This matters because it shows discretization approximations reflect the continuous spectral profile in a non-asymptotic way, with bounds that do not depend on how fine the grid becomes. The authors focus on the plunge region of eigenvalues away from 0 and 1, which measures local degrees of freedom. By proving uniform spectral deviation estimates, they demonstrate that theoretical localization properties remain observable in practical discrete computations such as Gabor multipliers.

Core claim

We study the eigenvalue profile of concentration operators, defined by multiplication with an indicator function followed by projection onto a reproducing kernel Hilbert space, and estimate the size of the plunge region consisting of eigenvalues away from 0 and 1. This provides a uniform notion of local degrees of freedom that holds for both continuous and discrete settings. Concretely, Gabor multipliers computed on sufficiently fine grids obey spectral deviation estimates matching those of the short-time Fourier transform, with bounds independent of the discretization step.

What carries the argument

Concentration operators (multiplication by an indicator function followed by orthogonal projection onto the reproducing kernel Hilbert space) whose eigenvalue distribution controls the plunge region between 0 and 1.

If this is right

The localization properties of the short-time Fourier transform become directly observable in finite-grid discrete computations without requiring asymptotic limits.
Discretization schemes for reproducing kernel spaces yield spectral profiles that match their continuous counterparts with error controlled uniformly in the grid step.
The measure of local degrees of freedom given by the plunge region size applies equally to both continuous and discrete concentration operators.
Theoretical bounds on spectral deviation can be used to certify the accuracy of practical Gabor multiplier implementations.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Similar uniform transfer of plunge-region estimates may apply to other time-frequency operators discretized on regular grids beyond the Gabor case.
The approach could guide the choice of grid density in applications by providing explicit fineness thresholds tied to kernel regularity.
It suggests that eigenvalue computations on discrete grids can serve as reliable proxies for continuous spectral analysis in settings where direct continuous computation is infeasible.

Load-bearing premise

The discretization grid must be sufficiently fine and the reproducing kernel together with the window function must satisfy regularity conditions that permit the spectral estimates to transfer uniformly from the continuous to the discrete setting.

What would settle it

Compute the eigenvalues of a Gabor multiplier on successively finer grids and observe whether the number of eigenvalues in the plunge region remains bounded by the continuous STFT estimate or grows without bound as the grid refines.

read the original abstract

We study the eigenvalue profile of concentration operators (multiplication by an indicator function followed by projection) acting on reproducing kernel Hilbert spaces. The spectral profile of such operators provides a useful notion of local degrees of freedom. We formalize this idea by estimating the number of eigenvalues that lie away from 0 and 1, commonly referred to as the plunge region. Our main motivation is to treat discrete and continuous settings simultaneously and uniformly, and to be able to argue that approximations arising from discretization schemes reflect, in a non-asymptotic sense, the spectral profile of their continuous counterparts. As a case in point, we show that Gabor multipliers computed on sufficiently fine grids obey spectral deviation estimates similar to those available for the short-time Fourier transform (STFT) with bounds that are uniform in the discretization step. Concretely, this means that the theoretical localization properties of the STFT are observable in practice.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper gives uniform spectral deviation bounds for Gabor multipliers on fine grids that match continuous STFT estimates without depending on the step size.

read the letter

The main thing to know is that this work shows spectral deviation estimates for concentration operators on RKHS that carry over uniformly from the continuous short-time Fourier transform to discrete Gabor multipliers on fine grids, with bounds independent of the discretization step. This uniformity is positioned as the key feature that lets you argue the discrete approximations reflect the continuous spectral profile in a non-asymptotic way. They do a good job framing the plunge region as a practical count of local degrees of freedom and motivating why you need non-asymptotic control so that discretization doesn't mess up the theoretical picture. The case for Gabor multipliers is presented as an example where the uniformity holds under suitable regularity on the kernel and window. This seems like a direct response to the need for results that justify numerical implementations without extra error terms depending on the grid. The abstract indicates they achieve this by imposing conditions that allow non-asymptotic control of the discretization error. The math appears grounded in standard operator theory for these spaces, and the abstract positions the uniformity as the novel part compared to prior continuous estimates. No signs of circular reasoning or reliance on fitted parameters from the description. A minor soft spot is the lack of detail in the abstract on exactly what regularity is required or how the discretization error is controlled to keep bounds uniform. Without seeing the proofs, it's tough to gauge if the conditions are mild or if they restrict the result to narrow classes of windows. If the paper provides concrete examples or explicit bounds, that would help. The stress test didn't flag inconsistencies, so probably the construction works as stated. Overall, this is for researchers in time-frequency analysis who work with both theory and numerics. It could be useful for anyone implementing localization operators and wanting to know the spectral profile stays reliable under discretization. A reader who follows work on concentration operators or Gabor systems would find the uniform bounds relevant. I would recommend sending it to peer review. The uniformity result looks like it could be a solid technical contribution worth checking in detail, even if the broader impact stays within the subfield.

Referee Report

2 major / 2 minor

Summary. The manuscript studies the eigenvalue distribution of concentration operators (multiplication by an indicator followed by orthogonal projection) on reproducing kernel Hilbert spaces. It focuses on bounding the size of the 'plunge region' (eigenvalues strictly between 0 and 1) and establishes non-asymptotic estimates that apply uniformly to both continuous and discrete settings. The main result shows that Gabor multipliers on sufficiently fine grids inherit spectral deviation bounds from the continuous short-time Fourier transform (STFT), with the bounds independent of the discretization step once the grid satisfies a regularity-dependent fineness condition.

Significance. If the uniformity claims hold, the work supplies a rigorous justification for transferring continuous localization results to discrete computations in time-frequency analysis without asymptotic loss of accuracy. The simultaneous treatment of continuous and discrete cases under explicit regularity assumptions on the kernel and window is a clear strength, as is the emphasis on non-asymptotic control rather than limiting arguments.

major comments (2)

[§4, Theorem 4.1] §4, Theorem 4.1: The proof of grid-independent bounds proceeds by controlling the discretization error via the regularity of the reproducing kernel; however, the constant C in the fineness condition (grid spacing < C) is stated to depend on the kernel's smoothness parameters, yet no explicit dependence or computable bound is derived, which weakens the claim of practical observability of the continuous spectral profile.
[§5.2, Eq. (5.7)] §5.2, Eq. (5.7): The operator-norm estimate for the difference between the continuous concentration operator and its discrete Gabor-multiplier counterpart relies on an integral remainder that assumes the window function belongs to a Sobolev-type space; this assumption is not verified for the standard Gaussian window used in the numerical examples, leaving a gap between the theorem statement and the reported experiments.

minor comments (2)

[§2] The notation for the plunge-region cardinality (often denoted N(ε)) is introduced in §2 but used with varying ε-thresholds in later sections without a uniform definition; a single notational convention would improve readability.
[Figure 3] Figure 3 caption refers to 'eigenvalue profiles for three grid sizes' but does not indicate the precise grid spacings or the value of the plunge threshold ε used to count outliers; adding these values would make the figure self-contained.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments. The positive assessment of the uniformity between continuous and discrete settings is appreciated. We address each major comment below and indicate the revisions planned.

read point-by-point responses

Referee: [§4, Theorem 4.1] §4, Theorem 4.1: The proof of grid-independent bounds proceeds by controlling the discretization error via the regularity of the reproducing kernel; however, the constant C in the fineness condition (grid spacing < C) is stated to depend on the kernel's smoothness parameters, yet no explicit dependence or computable bound is derived, which weakens the claim of practical observability of the continuous spectral profile.

Authors: We agree that an explicit expression for the constant C would enhance the practical utility of Theorem 4.1. The proof already tracks the dependence on the kernel's Sobolev norms through the discretization error estimates; we will extract and state the explicit form of C (in terms of the kernel regularity parameters and other constants from the proof) in the revised version of the theorem and its proof. revision: yes
Referee: [§5.2, Eq. (5.7)] §5.2, Eq. (5.7): The operator-norm estimate for the difference between the continuous concentration operator and its discrete Gabor-multiplier counterpart relies on an integral remainder that assumes the window function belongs to a Sobolev-type space; this assumption is not verified for the standard Gaussian window used in the numerical examples, leaving a gap between the theorem statement and the reported experiments.

Authors: The standard Gaussian window is C^∞ and therefore belongs to every Sobolev space H^s(ℝ) for s > 0. The integral remainder estimate in (5.7) therefore applies directly to the Gaussian window employed in the numerical examples. We will add a short clarifying remark in §5.2 to record this fact and close the gap between the theorem and the experiments. revision: partial

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper's central claims rest on operator-theoretic estimates for concentration operators in reproducing kernel Hilbert spaces, transferring spectral deviation bounds from continuous STFT settings to discrete Gabor multipliers under explicit regularity assumptions on kernels and windows. These assumptions enable non-asymptotic control of discretization error without reducing any prediction or eigenvalue count to a fitted parameter or self-referential definition. No load-bearing step invokes a self-citation chain, uniqueness theorem from prior author work, or ansatz smuggled via citation; the derivation remains independent of the target results and is externally falsifiable via the stated regularity conditions. The abstract and motivation explicitly frame the work as producing grid-independent bounds once grids are sufficiently fine, confirming self-contained operator analysis rather than circular reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard axioms of reproducing kernel Hilbert spaces, boundedness of projection operators, and properties of the short-time Fourier transform; no free parameters, ad-hoc axioms, or invented entities are introduced in the abstract.

axioms (2)

standard math Reproducing kernel Hilbert spaces are well-defined Hilbert spaces with continuous point evaluations
Invoked implicitly when defining concentration operators via projection onto the RKHS
standard math Concentration operators are bounded self-adjoint operators on the RKHS
Required for the eigenvalue profile to be well-defined and real

pith-pipeline@v0.9.0 · 5455 in / 1334 out tokens · 56064 ms · 2026-05-15T13:32:22.292238+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel (J-cost uniqueness) unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Theorem 2.2: |#{λ ∈ σ(T_Ω) : λ > δ} − μ(Ω)| ≲ ∇(Ω) · inf_s (τ N(s))^{γ/s} (log* …)^{1−γ/s} for s ≥ γ, with N(s) the dyadic decay integral of ||K(x,x')||_{S²}.
IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Corollary 2.4 and Theorem 8.3: uniform spectral deviation for Gabor multipliers M_{g,N,Ω} on fine lattices, bounds independent of N once N ≥ N_0(g,β).

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

55 extracted references · 55 canonical work pages · 1 internal anchor

[1]

Ambrosio, M

L. Ambrosio, M. Miranda Jr., and D. Pallara. Special functions of bounded variation in doubling met- ric measure spaces. InCalculus of Variations: Topics from the Mathematical Heritage of E. De Giorgi, volume 14 ofQuad. Mat., pages 1–45. Dept. Math., Seconda Univ. Napoli, Caserta, 2004

work page 2004
[2]

Ameur, F

Y. Ameur, F. Marceca, and J. L. Romero. Gaussian beta ensembles: the perfect freezing transition and its characterization in terms of Beurling-Landau densities.Ann. Inst. H. Poincar´ e Probab. Statist., 62(1):296– 327, 2026

work page 2026
[3]

Ameur and J

Y. Ameur and J. L. Romero. The planar low temperature Coulomb gas: separation and equidistribution. Rev. Mat. Iberoam., 39(2):611–648, 2023

work page 2023
[4]

B´ enyi and K

´A. B´ enyi and K. A. Okoudjou.Modulation Spaces: With Applications to Pseudodifferential Operators and Nonlinear Schr¨ odinger Equations. Applied and Numerical Harmonic Analysis. Birkh¨ auser Basel, 2020

work page 2020
[5]

Bj¨ orn and J

A. Bj¨ orn and J. Bj¨ orn.Nonlinear Potential Theory on Metric Spaces, volume 17 ofEMS Tracts in Mathe- matics. European Mathematical Society (EMS), Z¨ urich, 2011

work page 2011
[6]

S. G. Bobkov and F. G¨ otze. Discrete isoperimetric and Poincar´ e-type inequalities.Probab. Theory Related Fields, 114(2):245–277, 1999

work page 1999
[7]

Bonami, P

A. Bonami, P. Jaming, and A. Karoui. Non-asymptotic behavior of the spectrum of the sinc-kernel operator and related applications.J. Math. Phys., 62(3):Paper No. 033511, 2021. 34 F. MARCECA, J. L. ROMERO, M. SPECKBACHER, AND L. V ALENTINI

work page 2021
[8]

D. G. Caraballo. Areas of level sets of distance functions induced by asymmetric norms.Pacific J. Math., 218(1):37–52, 2005

work page 2005
[9]

J. Cho. A characterization of Gelfand-Shilov space based on Wigner distribution.Commun. Korean Math. Soc., 14(4):761–767, 1999

work page 1999
[10]

Chung, S.-Y

J. Chung, S.-Y. Chung, and D. Kim. Characterizations of the Gelfand-Shilov spaces via Fourier transforms. Proc. Amer. Math. Soc., 124(7):2101–2108, 1996

work page 1996
[11]

L. A. Coburn. The Bargmann isometry and Gabor-Daubechies wavelet localization operators. InSystems, approximation, singular integral operators, and related topics (Bordeaux, 2000), volume 129 ofOper. Theory Adv. Appl., pages 169–178. Birkh¨ auser, Basel, 2001

work page 2000
[12]

R. R. Coifman and Y. Meyer. Remarques sur l’analyse de Fourier ` a fenˆ etre.C. R. Acad. Sci. Paris S´ er. I Math., 312(3):259–261, 1991

work page 1991
[13]

Daubechies

I. Daubechies. Time-frequency localization operators: a geometric phase space approach.IEEE Trans. Inform. Theory, 34(4):605–612, 1988

work page 1988
[14]

M. A. Davenport and M. B. Wakin. Compressive sensing of analog signals using discrete prolate spheroidal sequences.Appl. Comput. Harmon. Anal., 33(3):438–472, 2012

work page 2012
[15]

David and S

G. David and S. Semmes.Analysis of and on Uniformly Rectifiable Sets, volume 38 ofMathematical Surveys and Monographs. American Mathematical Society, Providence, RI, 1993

work page 1993
[16]

De Mari, H

F. De Mari, H. G. Feichtinger, and K. Nowak. Uniform eigenvalue estimates for time-frequency localization operators.J. London Math. Soc. (2), 65(3):720–732, 2002

work page 2002
[17]

D¨ orfler, F

M. D¨ orfler, F. Luef, H. McNulty, and E. Skrettingland. Time-frequency analysis and coorbit spaces of operators.J. Math. Anal. Appl., 534:128058, 2024

work page 2024
[18]

Engliˇ s

M. Engliˇ s. Toeplitz operators and localization operators.Trans. Amer. Math. Soc., 361(2):1039–1052, 2009

work page 2009
[19]

I. M. Gelfand and G. E. Shilov.Generalized Functions. Vol. 2. AMS Chelsea Publishing, Providence, RI, 2016. Spaces of fundamental and generalized functions, Translated from the 1958 Russian original [MR0106409] by M. D. Friedman, A. Feinstein and C. P. Peltzer, Reprint of the 1968 English translation [MR0230128]

work page 2016
[20]

Gr¨ ochenig.Foundations of Time-Frequency Analysis

K. Gr¨ ochenig.Foundations of Time-Frequency Analysis. Applied and Numerical Harmonic Analysis. Birkh¨ auser Boston, Inc., Boston, MA, 2001

work page 2001
[21]

Gr¨ ochenig and M

K. Gr¨ ochenig and M. Leinert. Wiener’s lemma for twisted convolution and Gabor frames.J. Amer. Math. Soc., 17(1), 2003

work page 2003
[22]

Gr¨ ochenig and G

K. Gr¨ ochenig and G. Zimmermann. Spaces of test functions via the STFT.J. Funct. Spaces Appl., 2(1):25– 53, 2004

work page 2004
[23]

Halvdansson

S. Halvdansson. Empirical plunge profiles of time-frequency localization operators.Appl. Comput. Harmon. Anal., 81:101825, 2026

work page 2026
[24]

Heinonen and P

J. Heinonen and P. Koskela. Quasiconformal maps in metric spaces with controlled geometry.Acta Math., 181(1):1–61, 1998

work page 1998
[25]

Hern´ andez and G

E. Hern´ andez and G. Weiss.A First Course on Wavelets. Studies in Advanced Mathematics. CRC Press, Boca Raton, FL, 1996

work page 1996
[26]

Hughes, A

K. Hughes, A. Israel, and A. Mayeli. On the eigenvalue distribution of spatio-spectral limiting operators in higher dimensions, II.J. Fourier Anal. Appl., 31(51), 2025

work page 2025
[27]

Hughes, A

K. Hughes, A. Israel, and A. Mayeli. Wave packets and eigenvalue estimates for limiting operators on the disk.arXiv:2601.21224, 2026

work page arXiv 2026
[28]

A. Israel. The eigenvalue distribution of time-frequency localization operators.arXiv:1502.04404, 2015

work page internal anchor Pith review Pith/arXiv arXiv 2015
[29]

Israel and A

A. Israel and A. Mayeli. On the eigenvalue distribution of spatio-spectral limiting operators in higher dimensions.Appl. Comput. Harmon. Anal., 70:Paper No. 101620, 2024

work page 2024
[30]

Karnik, J

S. Karnik, J. Romberg, and M. A. Davenport. Improved bounds for the eigenvalues of prolate spheroidal wave functions and discrete prolate spheroidal sequences.Appl. Comput. Harmon. Anal., 55:97–128, 2021

work page 2021
[31]

A. Kulikov. Exponential lower bound for the eigenvalues of the time-frequency localization operator before the plunge region.Appl. Comput. Harmon. Anal., 71:Paper No. 101639, 8, 2024

work page 2024
[32]

A. Kulikov. Sharp estimates for eigenvalues of localization operators before the plunge region. arXiv:2603.07407, 2026

work page arXiv 2026
[33]

Kulikov and M

A. Kulikov and M. D. Larsen. Sharp estimates for eigenvalues of localization operators with applications to area laws.arXiv preprint: 2603.23832, 2026. SPECTRAL DEVIATION OF CONCENTRATION OPERATORS ON RKHS 35

work page arXiv 2026
[34]

P. Lahti. A Federer-style characterization of sets of finite perimeter on metric spaces.Calc. Var. Partial Differential Equations, 56(5):Paper No. 150, 2017

work page 2017
[35]

H. J. Landau. On Szeg˝ o’s eigenvalue distribution theorem and non-Hermitian kernels.J. Analyse Math., 28:335–357, 1975

work page 1975
[36]

H. J. Landau and H. O. Pollak. Prolate spheroidal wave functions, Fourier analysis and uncertainty. II. Bell System Tech. J., 40:65–84, 1961

work page 1961
[37]

H. J. Landau and H. O. Pollak. Prolate spheroidal wave functions, Fourier analysis and uncertainty. III. The dimension of the space of essentially time- and band-limited signals.Bell System Tech. J., 41:1295–1336, 1962

work page 1962
[38]

H. J. Landau and H. Widom. Eigenvalue distribution of time and frequency limiting.J. Math. Anal. Appl., 77(2):469–481, 1980

work page 1980
[39]

Luef and E

F. Luef and E. Skrettingland. On accumulated Cohen’s class distributions and mixed-state localization operators.Constr. Approx., 52:31–64, 2020

work page 2020
[40]

Marceca and J

F. Marceca and J. L. Romero. Spectral deviation of concentration operators for the short-time Fourier transform.Studia Math., 270(2):145–173, 2023

work page 2023
[41]

Marceca and J

F. Marceca and J. L. Romero. Improved discrepancy for the planar Coulomb gas at low temperatures.J. Anal. Math., 157(1):113–153, 2025

work page 2025
[42]

Marceca, J

F. Marceca, J. L. Romero, and M. Speckbacher. Eigenvalue estimates for Fourier concentration operators on two domains.Arch. Ration. Mech. Anal., 248(3):Paper No. 35, 2024

work page 2024
[43]

J. M. Maz´ on, J. D. Rossi, and J. J. Toledo.Nonlocal Perimeter, Curvature and Minimal Surfaces for Measurable Sets. Frontiers in Mathematics. Birkh¨ auser/Springer, Cham, 2019

work page 2019
[44]

Miranda Jr

M. Miranda Jr. Functions of bounded variation on “good” metric spaces.J. Math. Pures Appl. (9), 82(8):975–1004, 2003

work page 2003
[45]

J. P. Oldfield. Two-term Szeg˝ o theorem for generalised anti-Wick operators.J. Spectr. Theory, 5(4):751– 781, 2015

work page 2015
[46]

A. Osipov. Certain upper bounds on the eigenvalues associated with prolate spheroidal wave functions. Appl. Comput. Harmon. Anal., 35(2):309–340, 2013

work page 2013
[47]

Paulsen and M

V. Paulsen and M. Raghupathi.An Introduction to the Theory of Reproducing Kernel Hilbert Spaces. Cambridge University Press, 2016

work page 2016
[48]

Skrettingland

E. Skrettingland. Equivalent norms for modulation spaces from positive Cohen’s class distributions.J. Fourier Anal. Appl., 28(2), 2022

work page 2022
[49]

Slepian and H

D. Slepian and H. O. Pollak. Prolate spheroidal wave functions, Fourier analysis and uncertainty. I.Bell System Tech. J., 40:43–63, 1961

work page 1961
[50]

A. V. Sobolev. Pseudo-differential operators with discontinuous symbols: Widom’s conjecture.Mem. Amer. Math. Soc., 222(1043):vi+104, 2013

work page 2013
[51]

Soria and P

J. Soria and P. Tradacete. The least doubling constant of a metric measure space.Ann. Acad. Sci. Fenn. Math., 44(2):1015–1030, 2019

work page 2019
[52]

D. Thomson. Spectrum estimation and harmonic analysis.Proceedings of the IEEE, 70(9):1055–1096, 1982

work page 1982
[53]

Valentini

L. Valentini. Spectral deviation of Toeplitz operators acting on reproducing kernel Hilbert spaces. Master’s thesis, Universit` a degli studi di Genova, 2024

work page 2024
[54]

H. Widom. On a class of integral operators on a half-space with discontinuous symbol.J. Funct. Anal., 88(1):166–193, 1990

work page 1990
[55]

Zemen and C

T. Zemen and C. F. Mecklenbr¨ auker. Time-variant channel estimation using discrete prolate spheroidal sequences.IEEE Trans. Signal Process., 53(9):3597–3607, 2005. 36 F. MARCECA, J. L. ROMERO, M. SPECKBACHER, AND L. V ALENTINI (F. M.)Department of Mathematics, University College London, Gower Street, London WC1E 6BT, UK Email address:f.marceca@ucl.ac.uk ...

work page 2005