pith. sign in

arxiv: 2009.05029 · v5 · submitted 2020-09-10 · 🧮 math.FA

Phase retrieval of bandlimited functions for the wavelet transform

Rima Alaifari , Francesca Bartolucci , Matthias Wellershoff This is my paper

Pith reviewed 2026-05-24 14:31 UTC · model grok-4.3

classification 🧮 math.FA
keywords wavelet phase retrievaluniquenessbandlimited signalsvanishing momentsmagnitude measurementssignal recoveryfunctional analysis
0
0 comments X

The pith

Any wavelet with finitely many vanishing moments uniquely recovers real-valued bandlimited signals from the magnitude of its transform, up to global sign.

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

The paper examines phase retrieval for wavelet transforms, meaning the recovery of a signal solely from the absolute values of its wavelet coefficients. It proves a uniqueness result: real-valued bandlimited signals can be recovered uniquely up to sign flip when the analyzing wavelet has only finitely many vanishing moments. This matters because many signals in applications are bandlimited, and the result removes stricter requirements on the wavelet that appeared in earlier work. The authors also give the first uniqueness statements for sampled measurements, including with complex wavelets and for the Cauchy wavelet.

Core claim

Any wavelet with finitely many vanishing moments allows for the unique recovery of real-valued bandlimited signals up to global sign from the absolute values of their wavelet transforms. The paper additionally establishes the first uniqueness result for sampled wavelet phase retrieval when the wavelets may be complex-valued, together with a uniqueness result for phase retrieval from sampled Cauchy wavelet transform measurements.

What carries the argument

The continuous wavelet transform of real bandlimited functions, with the finite-vanishing-moments condition on the wavelet serving as the mechanism that forces uniqueness up to sign.

If this is right

  • Phase retrieval becomes possible for every wavelet satisfying the finite-vanishing-moments condition when the signal is real and bandlimited.
  • Sampled wavelet measurements suffice for unique recovery even when the wavelet is complex-valued.
  • Sampled Cauchy wavelet measurements likewise determine the underlying real bandlimited signal up to sign.

Where Pith is reading between the lines

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

  • The bandlimited restriction may be removable in future work by combining this result with density arguments.
  • The uniqueness could guide the design of reconstruction algorithms that alternate between wavelet magnitude constraints and bandlimit projections.
  • Similar finite-moment conditions might yield uniqueness statements for other integral transforms on bandlimited domains.

Load-bearing premise

The signals are real-valued and bandlimited and the wavelet has finitely many vanishing moments.

What would settle it

Two distinct real bandlimited functions, not differing by a global sign, whose wavelet transforms have identical absolute values for some wavelet with finitely many vanishing moments.

Figures

Figures reproduced from arXiv: 2009.05029 by Francesca Bartolucci, Matthias Wellershoff, Rima Alaifari.

Figure 1
Figure 1. Figure 1: The Morlet wavelet ψ for ξ0 = 5 in time and frequency representation. Observe that the Fourier transform of ψ is not identically zero on the negative frequencies but it is numerically small. in [20, 21]. We also refer to [16] for a concise presentation. Let ξ0, β ∈ R. The chirplet is defined by windowing a linear chirp with a Gaussian: ψ(x) = e i(ξ0+βx/2)x e −x 2/2 + η(x). Again, the corrective term η is a… view at source ↗
Figure 2
Figure 2. Figure 2: The linear-chirp wavelet for β = 1 and ξ0 = 5 in time and frequency representa￾tion. 1 2 3 4 5 6 7 8 9 10 −0.1 0.1 0.2 0.3 0.4 0.5 ψb [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The Cauchy wavelet ψb(ξ) = ξ 2 e −ξ1ξ>0. measurements |Wψf| for progressive wavelets ψ (see Remark 12 in Section 3). It does, however, raise the following question: (Q) Is there a class of signals which can be recovered (up to global phase) from wavelet transform magnitude measurements with progressive mother wavelets? In general, this question is hard to answer. An elegant partial answer is however given … view at source ↗
read the original abstract

We study the recovery of square-integrable signals from the absolute values of their wavelet transforms, also called wavelet phase retrieval. We present a new uniqueness result for wavelet phase retrieval. To be precise, we show that any wavelet with finitely many vanishing moments allows for the unique recovery of real-valued bandlimited signals up to global sign. Additionally, we present the first uniqueness result for sampled wavelet phase retrieval in which the underlying wavelets are allowed to be complex-valued and we present a uniqueness result for phase retrieval from sampled Cauchy wavelet transform measurements.

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.

Referee Report

2 major / 2 minor

Summary. The manuscript studies phase retrieval from the modulus of the continuous wavelet transform. Its central result asserts that any wavelet possessing finitely many vanishing moments permits unique recovery (up to global sign) of real-valued bandlimited signals. Additional results are stated for uniqueness from sampled wavelet measurements when the wavelet may be complex-valued and for sampled Cauchy-wavelet measurements.

Significance. If the uniqueness statements are correct, the work supplies the first general uniqueness theorem for wavelet phase retrieval that requires only finite vanishing moments rather than stronger non-vanishing conditions on the Fourier transform of the wavelet. This would enlarge the set of admissible wavelets for theoretical and numerical phase-retrieval studies and would complement existing results that rely on analyticity or specific wavelet families.

major comments (2)
  1. [Abstract / §1] Abstract and §1 (central claim): the statement that 'any wavelet with finitely many vanishing moments' yields uniqueness for real bandlimited signals is not obviously compatible with the fact that vanishing moments control only the jet of hat{psi} at omega=0. The manuscript must either (a) prove that no other zeros of hat{psi} can occur inside the band of the signal or (b) add an auxiliary hypothesis that hat{psi}(omega) != 0 for omega in the support of the signal spectrum. Without one of these, the uniqueness claim is at risk of counter-examples consisting of signals whose spectra lie in a zero of hat{psi}.
  2. [§3] §3 (proof of the main uniqueness theorem): the argument must be checked for an explicit non-vanishing assumption on hat{psi} over the frequency interval of interest. If the proof proceeds solely from finite vanishing moments without controlling other zeros, the step that recovers the Fourier transform from the modulus of the wavelet transform is incomplete.
minor comments (2)
  1. [§2] Notation for the continuous wavelet transform and its sampled version should be introduced with a single consistent symbol set rather than switching between W_psi f and CWT notations.
  2. [§4] The statement of the sampled Cauchy-wavelet result should clarify whether the sampling grid is uniform or adapted to the scale parameter.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments. We address the major comments point by point below.

read point-by-point responses

  1. Referee: [Abstract / §1] Abstract and §1 (central claim): the statement that 'any wavelet with finitely many vanishing moments' yields uniqueness for real bandlimited signals is not obviously compatible with the fact that vanishing moments control only the jet of hat{psi} at omega=0. The manuscript must either (a) prove that no other zeros of hat{psi} can occur inside the band of the signal spectrum or (b) add an auxiliary hypothesis that hat{psi}(omega) != 0 for omega in the support of the signal spectrum. Without one of these, the uniqueness claim is at risk of counter-examples consisting of signals whose spectra lie in a zero of hat{psi}.

    Authors: We agree that the phrasing in the abstract and §1 requires clarification. Finite vanishing moments ensure admissibility by controlling the behavior at zero but do not rule out other zeros of hat{psi}. The uniqueness proof relies on hat{psi} being non-vanishing on the frequency support of the bandlimited signal to recover the Fourier transform from the modulus. We will revise the abstract and §1 to add the explicit auxiliary hypothesis that hat{psi}(omega) != 0 for omega in the support of the signal spectrum. This resolves the concern about potential counterexamples. revision: yes

  2. Referee: [§3] §3 (proof of the main uniqueness theorem): the argument must be checked for an explicit non-vanishing assumption on hat{psi} over the frequency interval of interest. If the proof proceeds solely from finite vanishing moments without controlling other zeros, the step that recovers the Fourier transform from the modulus of the wavelet transform is incomplete.

    Authors: The proof in §3 does use the non-vanishing of hat{psi} over the frequency interval to justify the inversion step from the modulus. This assumption is implicit in the recovery but not stated explicitly in the theorem or proof. We will revise §3 to state the non-vanishing hypothesis explicitly and reference it in the relevant steps. The finite vanishing moments are used separately to handle the low-frequency behavior. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation rests on direct mathematical arguments

full rationale

The paper states a uniqueness theorem for wavelet phase retrieval of real bandlimited signals using wavelets with finitely many vanishing moments. The abstract and provided excerpts present this as a new result derived from properties of the wavelet transform and bandlimitedness, without reference to fitted parameters, self-definitional constructs, or load-bearing self-citations that reduce the claim to its inputs. The skeptic concern addresses potential gaps in the theorem's hypotheses (additional zeros in the Fourier transform) but does not identify any circular reduction in the derivation chain itself. This is a standard non-circular mathematical proof structure.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard assumptions from functional analysis and wavelet theory without introducing free parameters or new entities.

axioms (2)
  • standard math Wavelet transform is well-defined on L2 functions with the given vanishing moments property
    Invoked to establish the transform properties used in uniqueness proofs.
  • domain assumption Bandlimited real-valued functions form a suitable function class for recovery
    The uniqueness holds specifically under this restriction on the signals.

pith-pipeline@v0.9.0 · 5611 in / 1296 out tokens · 38909 ms · 2026-05-24T14:31:55.399660+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Sampling at twice the Nyquist rate in two frequency bins guarantees uniqueness in Gabor phase retrieval

    math.FA 2022-06 unverdicted novelty 7.0

    Bandlimited signals are uniquely recoverable up to global phase from Gabor magnitudes sampled at twice the Nyquist rate in two frequency bins.

Reference graph

Works this paper leans on

28 extracted references · 28 canonical work pages · cited by 1 Pith paper · 1 internal anchor

  1. [1]

    Alaifari, I

    R. Alaifari, I. Daubechies, P. Grohs, and G. Thakur. Reco nstructing real-valued functions from unsigned coefficients with respect to wavelet and other frames. Journal of Fourier Analysis and Applications , 23(6):1480–1494, 2017

    work page 2017

  2. [2]

    Alaifari and P

    R. Alaifari and P. Grohs. Phase retrieval in the general s etting of continuous frames for Banach spaces. SIAM journal on mathematical analysis , 49(3):1895– 1911, 2017

    work page 1911

  3. [3]

    Alaifari and M

    R. Alaifari and M. Wellershoff. Uniqueness of STFT phase r etrieval for bandlim- ited functions. Applied and Computational Harmonic Analysis , 50:34–48, January 2021

    work page 2021

  4. [4]

    Balan, P

    R. Balan, P. Casazza, and D. Edidin. On signal reconstruc tion without phase. Applied and Computational Harmonic Analysis , 20(3):345–356, 2006

    work page 2006

  5. [5]

    J. Bruna. Sampling in complex and harmonic analysis. In European Congress of Mathematics, pages 225–246. Springer, 2001

    work page 2001

  6. [6]

    Audio Texture Synthesis with Scattering Moments

    J. Bruna and S. Mallat. Audio texture synthesis with scat tering moments. arXiv:1311.0407v1 [stat.AP], November 2013. 16

    work page internal anchor Pith review Pith/arXiv arXiv 2013

  7. [7]

    Cahill, P

    J. Cahill, P. Casazza, and I. Daubechies. Phase retrieva l in infinite-dimensional Hilbert spaces. Transactions of the American Mathematical Society, Series B, 3(3):63–76, 2016

    work page 2016

  8. [8]

    Daubechies

    I. Daubechies. Ten lectures on wavelets . SIAM, 1992

    work page 1992

  9. [9]

    Grafakos

    L. Grafakos. Classical Fourier analysis . Graduate Texts in Mathematics. Springer Science+Business Media, New York, third edition, 2014

    work page 2014

  10. [10]

    Gröchenig

    K. Gröchenig. Phase-retrieval in shift-invariant spa ces with Gaussian generator. Journal of Fourier Analysis and Applications , 26(3):1–15, 2020

    work page 2020

  11. [11]

    Gröchenig, J

    K. Gröchenig, J. L. Romero, and J. Stöckler. Sharp resul ts on sampling with derivatives in shift-invariant spaces and multi-window Gabor frames. Constructive Approximation, 51(1):1–25, 2020

    work page 2020

  12. [12]

    Grohs and L

    P. Grohs and L. Liehr. Injectivity of Gabor phase retrie val from lattice measure- ments. arXiv preprint arXiv:2008.07238 , 2020

    work page arXiv 2008

  13. [13]

    Grossman and J

    A. Grossman and J. Morlet. Decomposition of Hardy funct ions into square inte- grable wavelets of constant shape. SIAM J. Appl. Math. , 15:723–736, 1984

    work page 1984

  14. [14]

    C. Heil. A basis theory primer . Applied and Numerical Harmonic Analysis (ANHA). Springer Science+Business Media, expanded editio n, 2011

    work page 2011

  15. [15]

    Holighaus, G

    N. Holighaus, G. Koliander, Z. Průša, and L. D. Abreu. Ch aracterization of analytic wavelet transforms and a new phaseless reconstruc tion algorithm. IEEE Transactions on Signal Processing , 67(15):3894–3908, August 2019

    work page 2019

  16. [16]

    Holschneider

    M. Holschneider. Wavelets. Oxford Mathematical Monographs. The Clarendon Press, Oxford University Press, New York, 1995. An analysis tool, Oxford Science Publications

    work page 1995

  17. [17]

    P. Jaming. Uniqueness results in an extension of Pauli’ s phase retrieval problem. Applied and Computational Harmonic Analysis , 37(3):413–441, November 2014

    work page 2014

  18. [18]

    S. Mallat. A wavelet tour of signal processing . Academic Press, Inc., San Diego, CA, 1998

    work page 1998

  19. [19]

    Mallat and I

    S. Mallat and I. Waldspurger. Phase retrieval for the Ca uchy wavelet transform. Journal of Fourier Analysis and Applications , 21(6):1251–1309, 2015

    work page 2015

  20. [20]

    Mann and S

    S. Mann and S. Haykin. The chirplet transform: A general ization of Gabors logon transform. Vision Interface, 91:205–212, 1991

    work page 1991

  21. [21]

    Mann and S

    S. Mann and S. Haykin. The chirplet transform: Physical considerations. IEEE Transactions on Signal Processing , 43(11):2745–2761, 1995

    work page 1995

  22. [22]

    Reed and B

    M. Reed and B. Simon. Methods of modern mathematical physics. I . Academic Press, Inc. [Harcourt Brace Jovanovich, Publishers], New Y ork, second edition,

    work page

  23. [23]

    J. L. Romero. Sign retrieval in shift-invariant spaces with totally positive gener- ator. Journal of Fourier Analysis and Applications , 27(2):1–8, 2021

    work page 2021

  24. [24]

    W. Rudin. Functional analysis. International Series in Pure and Applied Mathe- matics. McGraw-Hill, Inc., New York, second edition, 1991

    work page 1991

  25. [25]

    K. Seip. Interpolation and sampling in spaces of analytic functions , volume 33 of University Lecture Series. American Mathematical Society, Providence, Rhode Island, 2004. 17

    work page 2004

  26. [26]

    G. Thakur. Reconstruction of bandlimited functions fr om unsigned samples. Jour- nal of Fourier Analysis and Applications , 17(4):720–732, 2010

    work page 2010

  27. [27]

    Virtanen

    T. Virtanen. Monaural sound source separation by nonne gative matrix factor- ization with temporal continuity and sparseness criteria. IEEE Transactions on Audio, Speech, and Language Processing , 15(3):1066–1074, March 2007

    work page 2007

  28. [28]

    Waldspurger

    I. Waldspurger. Phase retrieval for wavelet transform s. IEEE Transactions on Information Theory, 63(5):2993–3009, 2017. 18

    work page 2017