$\texttt{raccoon}$: A Python package for removing wiggle artifacts in the JWST NIRSpec integral field spectroscopy

Anowar J. Shajib

REVIEW

$\texttt{raccoon}$: A Python package for removing wiggle artifacts in the JWST NIRSpec integral field spectroscopy

Anowar J. Shajib

Instrumentation and Methods for Astrophysics · Submitted 17 Jul 2025 · 2 references

Pith's verdict: A plausible, clearly written software paper for a real JWST NIRSpec artifact, but the central empirical premise—that aperture sums are wiggle-free—is untested and needs quantitative validation before the claims carry weight.

T0 · AI review

How this review was made: An AI referee (deepseek-v4-flash) read the paper on August 6, 2026 and answered a fixed list of questions about it. T0 is the starting level every paper gets. The levels describe how deeply Pith has checked the paper, never the authors.

T0 AI review An AI read the paper and answered a fixed list of questions.
T1 Audit Automatic checks that the references and files the paper relies on exist and match.
T2 Compute The paper's code or calculations re-run with the same results.
T3 Formal statement The claims written out in a formal math language a computer can read.
T4 Proof Those claims proved, and the proof checked by computer.

How Pith checks papers →

· 3 major objections · 4 minor · Ask about this paper · The compiled record · Read the full review →

Challenge this review Comment (0) Read on arXiv ↗

Cite ▾

arXiv ID2507.13341 (version 1)

Pith IDFUVIPUC5, Pith's permanent record of this paper and its review

CitationAnowar J. Shajib. "$\texttt{raccoon}$: A Python package for removing wiggle artifacts in the JWST NIRSpec integral field spectroscopy." arXiv:2507.13341 (2025). https://pith.science/paper/FUVIPUC5

Cite this reviewBibTeX for Pith's review of this paper

Claim paper

Why Pith says this

Highlight any text to comment on it.

Editor's note

read the letter

Quick take: this package addresses a real problem (resampling wiggles in JWST NIRSpec IFS at single-spaxel level), and the simultaneous forward fit of a chirp times a template is a genuine, if incremental, improvement over the subtract-then-fit routines of Perna et al. and WICKED. The math is straightforward and clearly stated. The paper gives an honest description of the method and the figures show plausible fits to one AGN spectrum.

What's new: the simultaneous fit across the full wavelength range, the multiplicative correction, and the pip-installable implementation. Those are real but modest contributions. The core empirical idea—using aperture-summed spectra as wiggle-free templates—is inherited from prior work; the novelty is in the implementation.

Soft spots, in order of seriousness:

1. No quantitative validation. The paper shows illustrative fits but no injected-wiggle recovery tests, no comparison with WICKED or Perna et al. on the same data, and no residual statistics after correction. For a paper whose claim is that it restores fidelity, this is the missing piece.

2. The stress-test concern is legitimate. Eq. 3 constructs T from the aperture sum C, and if C retains any wiggle W_C, the corrected spectrum inherits it. The author cites Figure 1 and prior work for washout, but does not quantify the residual as a function of aperture radius or source morphology. This is an empirical premise, not an internal contradiction, but it is load-bearing. The degeneracy between W and smooth variations in the template is real, though the chirp form and B-spline knot counts constrain it somewhat. I would want to see a test where wiggles are injected into a synthetic spectrum and recovered.

3. The paper's self-citation is fine—TDCOSMO usage is evidence of adoption—but the placeholder DOI and missing repository/archive links are a concrete deficiency. The package is not actually available for inspection via the paper.

The method is not circular in the objectionable sense; it is a self-calibration, a standard trick. But self-calibration requires validation to be credible, and that is missing.

Who this is for: anyone working with JWST NIRSpec IFS at spaxel scale, especially stellar kinematics and line measurements. A serious referee should engage with it. I would condition acceptance on adding quantitative validation and making code available. The central idea holds up as a plausible empirical correction; it just needs to be tested rather than asserted.

Referee report

Significance

If the method works as claimed, raccoon would address a real limitation of JWST NIRSpec IFS data: PSF undersampling introduces spaxel-level spectral artifacts that the standard pipeline does not remove, and these can bias kinematic and emission-line measurements. The paper is clearly written, the mathematical model is explicit, the least-squares framework is standard, and the code is publicly available and already cited in scientific applications. The multiplicative model is physically motivated, and the simultaneous fit of template and wiggle is a principled improvement over sequential subtraction approaches. However, the current evidence for the central claim is anecdotal: only one spectrum is shown, with no quantitative recovery test, no comparison against WICKED, and no measurement of the residual wiggle in the aperture-summed template that underpins the correction.

Major objections (3)

[Functionality, Figures 1–3] The central claim that raccoon 'restores the fidelity of the extracted spectra' is supported only by a visual demonstration on a single AGN spectrum. There is no quantitative validation: no injected-wiggle recovery test, no comparison to WICKED on the same spaxel, and no metric showing that the corrected spectrum is closer to a truth spectrum than the uncorrected one. Please include at least one controlled experiment, e.g., inject a known Eq. (1) wiggle into a spectrum believed wiggle-free and verify that the fitted W(λ) recovers the injection and that the corrected spectrum matches the input within noise, and/or a side-by-side comparison with WICKED.
[Eqs. (2)–(3), Functionality] The load-bearing premise that the aperture-summed spectrum C(λ) is effectively wiggle-free is not quantified. The manuscript cites Figure 1 and prior work for washout, but it does not report the residual wiggle amplitude in aperture sums as a function of aperture radius or source morphology. Because W(λ) multiplies T(λ), any residual periodic component in C(λ) is degenerate with the fitted chirp; the model can absorb it into W(λ), leaving the corrected spectrum D/W_model contaminated by the template residual. Please measure or bound this residual on real or simulated data and show that it is small relative to the single-spaxel wiggle amplitude.
[Figure 3 caption] The caption states that the plotted points are 'D(λ)/M(λ) based on the best-fit M(λ)', with M(λ)=W(λ)T(λ) by Eq. (2). With a good fit D≈M, this ratio is approximately unity, not the extracted wiggle signal. The ratio shown appears to be D divided by the template (or the model without W), or the figure text misstates the plotted quantity. This should be corrected, because the figure is the primary evidence for the wiggle model's fidelity.

Minor objections (4)

[Eq. (1)] The notation 'a1 sin2 ϕ(λ)' is ambiguous; it likely means a1 sin(2ϕ(λ)) rather than a1 sin^2(ϕ(λ)). Please clarify with explicit parentheses in the text and equation.
[Functionality, after Eq. (4)] The 'χ2_MAP' metric for knot selection is mentioned but never defined; please define it or provide a reference.
[Abstract and Functionality] The abstract states that uncertainties are propagated, but the paper does not describe how the fitted wiggle parameters and their covariances are propagated into the corrected spectra. A brief description or reference to the implementation would be useful.
[Statement of need and Functionality] The claim that raccoon is 'uniquely installable via the pip command' is too strong without checking the distribution methods of the cited comparison tools; please soften to 'can be installed via pip'.

How Pith checks papers → · Full audit of this paper →

The paper

The paper claims

raccoon claims that a NIRSpec spaxel spectrum is a chirp wiggle multiplied by a smooth template, and that fitting both at once and dividing out the wiggle restores the underlying spectrum.

The reading

raccoon claims that the wiggles corrupting JWST NIRSpec integral-field spaxel spectra are not noise to be smoothed away but a deterministic artifact that can be modeled and removed. It describes every single-spaxel spectrum as a chirp-like wiggle function multiplied by a smooth template built from spectra summed over a circular aperture around that spaxel, with an optional shell sum. The package fits this product to the data across the full wavelength range in one step, then divides out the fitted wiggle to produce a corrected spectrum. If the model is faithful, spaxel-level science such as stellar kinematics and line measurements no longer has to be sacrificed or aperture-averaged to escape the artifacts.

What carries the argument

The central object is the product model $M(\lambda) = W(\lambda) T(\lambda)$: a chirp wiggle $W$ times a wiggle-free template $T$. The wiggle is written as $W(\lambda) = 1 + A(\lambda)[\sin \phi + a_1 \sin^2\phi + a_2 \sin(3\phi)]$ with phase $\phi = \lambda k(\lambda) + \phi_0$, allowing wavelength-dependent amplitude and wavenumber plus asymmetric or sharpened peak shapes. The template is built as $T(\lambda) = c_1 C(\lambda) + c_2 S(\lambda) + c_3 \lambda^b + \sum_{n=0}^N p_n \lambda^n$, where $C$ is the circular-aperture-summed spectrum and $S$ an optional shell sum, making the template empirical and local. B-splines for $A(\lambda)$ and $k(\lambda)$, the linear coefficients, and the shape parameters are all solved by one $\chi^2$ minimization across the full spectrum, after which the fitted $W$ is divided out of the data.

Extended reading notes

Core claim

The central claim is that the artifact obeys the multiplicative model $D(\lambda) = W(\lambda) T(\lambda)$ (Eq. 2), with $W(\lambda)$ a parametrized chirp (Eq. 1) whose amplitude and wavenumber are smooth B-spline functions of wavelength, and $T(\lambda)$ a linear-template combination of the circular-aperture-summed spectrum, an optional shell-summed spectrum, and a power-law-plus-polynomial continuum. Fitting $M(\lambda) = W(\lambda) T(\lambda)$ to the observed single-spaxel spectrum by minimizing $\chi^2$ over the whole wavelength range yields the wiggle signal in one step, and dividing the data by the best-fit $W$ removes the artifacts while propagating the associated uncertainties. The paper argues that this simultaneous fit preserves wiggle signal that earlier subtract-then-fit methods lose, and that a multiplicative correction is physically motivated because the wiggles act multiplicatively on the spectrum.

Load-bearing premise

The argument depends on the aperture-summed spectrum being essentially free of wiggles and representative of the true single-spaxel spectrum; if the wiggles do not cancel in the sum, or if the true spectrum varies within the aperture in a way the linear template cannot capture, the corrected spectrum will carry artifacts of that mismatch.

Editorial extensions

If this is right

Corrected spaxel-level spectra retain the native NIRSpec spatial resolution instead of requiring large-aperture sums, so kinematics and line measurements can be made at the finest sampling the instrument provides.
Because the fit covers the entire wavelength range in one step, masked regions or gaps from outlier rejection do not break the wiggle estimate.
A user-set detection threshold means corrections are applied only to spaxels with significant wiggles, leaving already clean spectra untouched.
The same chirp-plus-template machinery can incorporate a shell spectrum to track line-shape changes across a galaxy, extending the correction to kinematically varying sources.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the multiplicative chirp model generalizes, the same approach could be adapted to other undersampled integral-field or dithered instruments, since resampling artifacts commonly act as wavelength-dependent multiplicative ripples.
A testable prediction is that injecting a synthetic chirp of known amplitude and phase into a wiggle-free spaxel spectrum should be recovered by raccoon, and the corrected output should match the original spectrum within noise.
When the aperture-summed template fails to cancel wiggles for compact or highly variable sources, the correction would imprint template mismatch; users could detect this by comparing corrected spectra produced from two different aperture radii.
The claimed advantage over subtract-then-fit methods implies that the extracted wiggle depends on the template family, so systematic checks across template choices would quantify the remaining model-dependent uncertainty.

Original abstract

$\texttt{raccoon}$ is a Python package for removing resampling noise - commonly referred to as "wiggles'' - from spaxel-level spectra in datacubes obtained from the JWST Near Infrared Spectrograph's (NIRSpec) integral field spectroscopy (IFS) mode. These wiggles arise as artifacts during resampling of the 2D raw data into 3D datacubes, due to the point spread function (PSF) being undersampled. The standard JWST data reduction pipeline does not correct for this noise. The wiggle artifacts can significantly degrade the scientific usability of the data, particularly at the spaxel level, undermining the exquisite spatial resolution of NIRSpec. $\texttt{raccoon}$ provides an empirical correction by modeling and removing these artifacts, thereby restoring the fidelity of the extracted spectra. $\texttt{raccoon}$ forward-models the wiggles as a chirp function impacting one or more template spectra that are directly fit to the original data across the entire wavelength range. The best-fit wiggle model is then used to clean the data while propagating the associated uncertainties.

The compiled record

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-23: 1 request, 5,690 input tokens.

What the abstract itself states, as probabilities: claims first 0.06 · states limitation 0.11 · human participants 0.01 · controlled comparison 0.10 · randomized assignment 0.01 · states quantitative result 0.04 · compares with prior results 0.36 · mentions code or data availability 0.19.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; show every role.

$\texttt{raccoon}$ is a Python package for removing resampling noise - commonly referred to as "wiggles'' - from spaxel-level spectra in datacubes obtained from the JWST Near Infrared Spectrograph's (NIRSpec) integral field spectroscopy (IFS) mode. These wiggles arise as artifacts during resampling of the 2D raw data into 3D datacubes, due to the point spread function (PSF) being undersampled. The standard JWST data reduction pipeline does not correct for this noise. The wiggle artifacts can significantly degrade the scientific usability of the data, particularly at the spaxel level, undermining the exquisite spatial resolution of NIRSpec. $\texttt{raccoon}$ provides an empirical correction by modeling and removing these artifacts, thereby restoring the fidelity of the extracted spectra. $\texttt{raccoon}$ forward-models the wiggles as a chirp function impacting one or more template spectra that are directly fit to the original data across the entire wavelength range. The best-fit wiggle model is then used to clean the data while propagating the associated uncertainties.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	$\texttt{raccoon}$ is a Python package for removing resampling noise - commonly referred to as "wiggles'' - from spaxel-level spectra in datacubes obtained from the JWST Near Infrared Spectrograph's (NIRSpec) integral field spectroscopy (IFS) mode.	contribution 1.00	0.49	0.22	asserted 1.00
s1	These wiggles arise as artifacts during resampling of the 2D raw data into 3D datacubes, due to the point spread function (PSF) being undersampled.	background 1.00	0.15	0.04	asserted 1.00
s2	The standard JWST data reduction pipeline does not correct for this noise.	gap 0.77	0.10	0.04	asserted 1.00
s3	The wiggle artifacts can significantly degrade the scientific usability of the data, particularly at the spaxel level, undermining the exquisite spatial resolution of NIRSpec.	background 0.93	0.18	0.09	asserted 0.70
s4	$\texttt{raccoon}$ provides an empirical correction by modeling and removing these artifacts, thereby restoring the fidelity of the extracted spectra.	contribution 0.96	0.81	0.53	asserted 1.00
s5	$\texttt{raccoon}$ forward-models the wiggles as a chirp function impacting one or more template spectra that are directly fit to the original data across the entire wavelength range.	method 1.00	0.80	0.12	asserted 1.00
s6	The best-fit wiggle model is then used to clean the data while propagating the associated uncertainties.	method 1.00	0.78	0.14	asserted 1.00

What kind of paper the classifier reports this is: method or tool 1.00 · evidence from analysis of existing data 0.53.

What was examined, and what was not

sources read

arXiv listing abstract v1

not examined

full text
figures and tables
references and citing sentences
relations to other papers

levels run

P abstract

not run

S sections
T body sentences
C citations
R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

Figures

**Figure 1.** Illustration of the wiggles in the single-spaxel spectrum (blue), which is a manifestation of the resampling noise in the standard-pipeline-reduced datacube due to PSF undersampling. The red spectrum shows the aperture-summed spectrum within a 4-spaxel radius, normalized to match its maximum to that of the single-spaxel one. The illustrated data are of an active galactic nucleus (AGN) from Perna et al. (2023). In raccoon, the wiggle is modeled as a sinusoidal chirp function W(λ) = 1 + A(λ)

**Figure 2.** Modeling of the full spectrum (blue in the top panel) based on the template spectra and the wiggles impacting it. The illustrated spectrum is the same one from

**Figure 3.** Illustration of the extracted and modeled wiggle signal. The illustrated data points represent D(λ)/M(λ) based on the best-fit M(λ), and the orange line illustrates the best-fit wiggle model W(λ). The grey points are rejected outliers using the false discovery rate method, and the blue points mark the wavelengths where the data were fit to the model. raccoon enables users to efficiently process multiple spaxels in a datacube, either across the entire field or within a specified spatial region. A user-defined detection threshold for the wiggle signal can be set, ensuring that…

Figures reproduced from arXiv:2507.13341 by the authors.

References

Works cited (2)

[1]

Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B (Methodological), 57 (1), 289–300. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x B¨ oker, T., Beck, T. L., Birkmann, S. M., Giardino, G., Keyes, C., Kumari, N., Muzerolle...

View on Pith arXiv 1995
[45]

https://doi.org/10.3847/1538-3881/acdddc Perna, M., Arribas, S., Marshall, M., D’Eugenio, F., ¨Ubler, H., Bunker, A., Charlot, S., Carniani, S., Jakobsen, P., Maiolino, R., Rodr ´ ıguez Del Pino, B., Willott, C. J., B¨ oker, T., Circosta, C., Cresci, G., Curti, M., Husemann, B., Kumari, N., Lamperti, I., . . . Scholtz, J. (2023). GA-NIFS: The ultra-dense,...

View on Pith arXiv 2023

Forward citations

Cited by 3 Pith papers

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

The M87 Black Hole Mass from JWST and Keck Stellar Kinematics Spanning 10 pc to 10 kpc
astro-ph.GA 2026-11 accept novelty 6.0 of 10

Triaxial orbit modeling of stellar kinematics from JWST and Keck yields M87's black hole mass as (5.38 ± 0.13 ± 0.22) × 10^9 solar masses.
The JWST Ultramassive Galaxy Sample -- I. Overview, Photometry, and Stellar Kinematics: Selecting Optimal Template Subsets via Non-Negative Matrix Factorization
astro-ph.GA 2026-09 conditional novelty 5.0 of 10

JWST observations of eight ultramassive galaxies yield photometric models and stellar kinematics, with a new NMF-based method for selecting optimal template subsets from synthetic libraries.
TDCOSMO XXX: Spatially resolved kinematics of the deflectors in time-delay lens systems B1608+656 and SDSSJ1206+4332 from JWST-NIRSpec observation
astro-ph.GA 2026-09 conditional novelty 5.0 of 10

First resolved stellar kinematics of the deflectors in B1608+656 and SDSS J1206+4332 from JWST-NIRSpec, including first redshifts and velocity dispersions of the two perturbing companion galaxies.

How the literature uses it

1 arXiv paper cite this in 2 sentences

Each citing sentence is taken from the citing paper's LaTeX source and typed by Pith's classifier. A label appears at probability 0.8 or more (0.95 for support), a contest at 0.7 or more; the quoted words are the citing authors'. How each label was checked · JSON

Papers citing it · 1 sentence without a confident label

Contests it 1 paper

Several public tools have been developed to remove these artifacts [ref] by forward modeling them with sinusoidal functions; however, these methods do not perform effectively for our data, likely due to the faint and noisy nature of the spectra, which makes analytical modeling difficult.

Spatially Resolved Physical Properties of Young Star Clusters and Star-forming Clumps in the Brightest z>6 Galaxy, the Strongly Lensed Cosmic Spear at z=6.2 · Generating Spectrophotometric Data Cube · contest 0.85

Discussion

Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.

[1] [1]

Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society: Series B (Methodological), 57 (1), 289–300. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x B¨ oker, T., Beck, T. L., Birkmann, S. M., Giardino, G., Keyes, C., Kumari, N., Muzerolle...

View on Pith arXiv 1995

[2] [45]

https://doi.org/10.3847/1538-3881/acdddc Perna, M., Arribas, S., Marshall, M., D’Eugenio, F., ¨Ubler, H., Bunker, A., Charlot, S., Carniani, S., Jakobsen, P., Maiolino, R., Rodr ´ ıguez Del Pino, B., Willott, C. J., B¨ oker, T., Circosta, C., Cresci, G., Curti, M., Husemann, B., Kumari, N., Lamperti, I., . . . Scholtz, J. (2023). GA-NIFS: The ultra-dense,...

View on Pith arXiv 2023