Pith. sign in

REVIEW 3 major objections 5 minor 23 references

The paper claims that reanalysing the full RAVE DR6 spectral sample with a two-component stellar model reveals 2,813 composite spectra from 2,605 likely double-lined binary systems, a much larger set than the 123 found previously.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review

2026-08-02 01:01 UTC pith:5H5UFM56

load-bearing objection A solid but contingent RAVE SB2 catalog: the method is recycled from earlier work, the external checks are convincing, and the unexplained wavelength-calibration cluster keeps the headline number from being fully reproducible. the 3 major comments →

arxiv 2607.14794 v1 pith:5H5UFM56 submitted 2026-07-16 astro-ph.SR

Search for double-line spectroscopic binaries in RAVE survey

Mikhail Kovalev (YNAO) This is my paper
classification astro-ph.SR PACS 97.80.Fk
keywords spectroscopic binariesSB2RAVE surveycomposite spectral fittingradial velocitiest-SNE classificationwavelength calibrationstellar spectra
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper applies a composite spectral fitting method to all 518,302 spectra of the RAVE survey's final data release, aiming to uncover double-lined spectroscopic binaries (SB2s) that were not a primary survey product. It reports 2,813 spectra, belonging to 2,605 unique targets, as SB2 candidates after combining three complementary selection criteria and visual inspection. This is a major increase over the 123 SB2 detections from an earlier RAVE release. If the candidates are largely real, they would provide a large sample for studying binary statistics, testing stellar evolution, and verifying orbital solutions from other surveys. The paper also flags a wavelength-calibration artifact affecting some spectra and identifies 13 potential triple-lined systems.

Core claim

The central discovery is that a significant fraction of RAVE DR6 spectra show composite-line signatures consistent with two stellar components. By fitting each spectrum with a single-star and a two-star model and selecting on (1) the inflated vsini of the single-star fit, (2) the fit-improvement ratio, and (3) the survey's own binary classification flag, the author constructs a catalog of 2,813 candidate SB2 spectra (2,605 unique targets). This catalog includes radial velocities, spectral parameters, and t-SNE coordinates, and is offered as a community resource for orbit verification and binary studies.

What carries the argument

The key machinery is the composite spectral model: a binary spectrum is built as the weighted sum of two Doppler-shifted synthetic single-star spectra, with weights set by the Planck function and a free luminosity-ratio parameter k. An improvement factor f_imp compares absolute residuals of the single-star and binary fits, and three threshold selections (vsini-based, primary-based, and the survey's P='b' flag) are then visually inspected to remove false positives.

Load-bearing premise

The assumption that the restricted synthetic grid (Teff 5000–15000 K, |[Fe/H]|≤0.3 dex, two vsini values, equal component metallicities) adequately represents all resolvable SB2s in RAVE; binaries outside these ranges would not be fit well and would be missed.

What would settle it

Check whether known SB2 systems with primary Teff below 5000 K or |[Fe/H]|>0.3 are present in RAVE DR6 but absent from the candidate list; such omissions would directly demonstrate grid incompleteness. Alternatively, obtain high-resolution follow-up spectroscopy for ~50 randomly selected candidates; if many show no line splitting, the visual-selection contamination rate is significant.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • If the candidates are genuine, the RAVE survey's SB2 yield grows from 123 to over 2,600, providing a large sample for binary population studies in the southern sky.
  • The measured radial velocities for the two components can be combined with astrometric and spectroscopic orbits from other surveys to identify triple systems or detect orbital motion over the decade between observations.
  • The wavelength-calibration artifact (a differential shift of the calcium triplet lines) is identified and clustered in t-SNE space, so future surveys can screen for it.
  • The method, being cheap (a week on a four-core computer), can be scaled to upcoming large spectroscopic surveys with similar wavelength coverage.

Where Pith is reading between the lines

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

  • The paper's completeness is bounded by its restricted model grid (Teff 5000–15000 K, |[Fe/H]|≤0.3, and only two vsini values); systems with cooler or metal-rich components may be systematically missed, so the true SB2 count could be higher.
  • Because the final filter is visual inspection, the catalog likely contains a small, unquantified fraction of false positives; a purely quantitative re-evaluation with a machine-learning classifier could sharpen the selection.
  • The 13 SB3 candidates are intriguing but unconfirmed; if even a subset is real, they would be a rare sample of triple-lined systems useful for dynamical studies.
  • The t-SNE clustering of candidates suggests that an automatic embedding-based classifier trained on this catalog could replace visual inspection for future releases.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper searches the full RAVE DR6 spectral archive (518,302 spectra) for double-line spectroscopic binary (SB2) candidates by fitting each spectrum with single-star and composite binary spectral models. Three complementary selection methods are used: a (Vsini)-based criterion, a primary-component selection based on improved chi-square and f_imp, and the RAVE DR6 automatic 'b' classification flag. After visual inspection, the author reports 2,813 composite spectra belonging to 2,605 unique targets. The catalog is cross-matched against literature and Gaia NSS, yielding 220 Gaia SB2 matches and 8 SB9 matches; a detailed case is made for SX Oph, HD 20784, and a chance-alignment pair. The paper also identifies a wavelength-calibration problem affecting a t-SNE cluster near coordinates (97,83) and reports 13 possible SB3 systems.

Significance. If the result holds, this is a substantial advance for RAVE binary science: the known SB2 content grows from 123 detections in DR2 to a catalog of thousands of candidates, with radial velocities that can be combined with Gaia and other surveys for orbit verification. The paper makes good use of external anchors — the SX Oph radial velocities tracking the ASAS-SN light curve and the 220 Gaia NSS SB2 matches are direct, falsifiable checks that argue for the reality of the selection. The catalog is machine-readable and includes t-SNE coordinates and fit parameters, which is a strength for reproducibility. The main caveats are that the paper does not quantify how the restricted model grid affects completeness, and it leaves the known wavelength-calibration contamination unquantified in the final sample.

major comments (3)
  1. [§3.3.5 and Table 2] The paper explicitly states that the S-shaped wavelength-calibration distortion makes binary models fit these spectra 'a bit better than single-star model, so they get selected as SB2 candidates,' and that 238 such spectra cluster near t-SNE coordinates (97,83). However, it is never stated whether these 238 spectra are excluded from the final 2,813-count catalog, and Table 2 has no flag for wavelength-calibration problems. Because the headline result is the number 2,813/2,605, this known systematic could inflate the sample by up to 8.5%. Please state how many of the 238 are retained in the final catalog, add an explicit flag to the table, and justify any retained cases or remove them.
  2. [§3.1, Eq. (2)–(3)] The model grid is restricted to Teff 5000–15000 K, |[Fe/H]|≤0.3 dex, equal metallicity for both components, and only two (Vsini) values. The paper does not quantify completeness loss for real binaries with components outside these ranges, e.g., cool M-type primaries, hot O/B secondaries, or metal-poor/rich pairs. Since all three selection channels in Table 1 depend on chi2 and f_imp from these models, the completeness of the 2,813 candidate count is unknown. A recovery-injection test or a coverage diagnostic showing the distribution of single-star best-fit parameters relative to grid edges would make the detection efficiency clearer; without it, the catalog cannot be interpreted as a full census.
  3. [§3.2 and Table 1] The final numbers 'after check' (2304, 885, 1218) and the final 2,813 spectra are obtained through visual inspection, but the visual decision procedure is not described. No examples of rejected spectra, no rejection criteria, and no code for reproducing the visual check are provided. This makes the final selection step non-transparent and not independently repeatable from the paper alone. Please either specify concrete criteria used during visual inspection or provide the full pre- and post-inspection lists in the catalog so that users can reproduce the filtering.
minor comments (5)
  1. [Abstract] The abstract says 'from the RAVE spectra' — likely should be 'from the RAVE survey' or 'from RAVE DR6 spectra'.
  2. [§3.2 vs Table 1] The text reports that 1220 spectra passed the DR6 'b' flag check, while Table 1 lists '1218 after check'. Please harmonize these numbers.
  3. [§3.3.2] The statement that the t-SNE map 'shows good agreement' is qualitative. A quantitative comparison, e.g., the overlap fraction between the selected candidates and the highlighted clusters, would strengthen the claim.
  4. [General] The paper uses the writing 'RA VE' (with space) throughout; standard usage is 'RAVE'. Also, 'All these three selections having 2813 spectra' in §3.2 is ungrammatical and should be revised.
  5. [Table 2] The catalog table lists 'SB2 index' as a confirmation of visual inspection, but no equivalent flag for the wavelength-calibration problem is present. Adding such a flag would also address the major concern above.

Circularity Check

0 steps flagged

No circular derivation: the 2813/2605 SB2 count is an empirical search output, and the self-citations are methodological lineage rather than load-bearing evidence.

full rationale

The paper's central number (2813 composite spectra / 2605 unique targets) is the measured result of a model-fitting and selection pipeline applied to 518,302 RAVE DR6 spectra; it is not a quantity derived from a fitted constant or from the selection thresholds themselves. The single-star and binary model equations (1)-(4) are stated explicitly, and the selection criteria in Table 1 are fixed quality cuts, not parameters fitted to the final count. The method is inherited from the author's earlier papers (Kovalev & Straumit 2022; Kovalev et al. 2022, 2024b), but this is methodological lineage rather than a self-citation chain that forces the result: the resulting catalog is checked against independent external data (Matijevic et al. 2010, Gaia NSS tables, SB9, ASAS-SN light curve of SX Oph, Gaia astrometry for J132346.4-675653). The t-SNE comparison is unsupervised and not constructed from the fit outputs. The one genuinely concerning passage is Section 3.3.5, where the author notes that spectra with an S-shaped wavelength-calibration distortion 'get selected as SB2 candidates' and identifies a 238-spectrum t-SNE cluster; the paper does not explicitly state whether these were removed from the 2813. However, this is a catalog contamination/completeness issue, not circular reduction: the count is not defined in terms of the flagged cluster, and no selection criterion is the inverse of the claimed result. Therefore no circular step is exhibited.

Axiom & Free-Parameter Ledger

4 free parameters · 5 axioms · 0 invented entities

The central numbers depend on (1) the accepted physics of the binary model (sum of two Planck-scaled spectra), (2) the coverage of the synthetic grid, (3) the hand-set thresholds in Table 1, and (4) the visual inspection step. These are inputs the reader must accept from prior literature or from the paper's stated choices; they are not derived within this manuscript.

free parameters (4)
  • Model grid metallicity restriction = |[Fe/H]| ≤ 0.3 dex
    Excludes metal-rich/metal-poor models to speed up fitting; real SB2s with components outside this range may be under-fitted and missed (Section 3.1).
  • Model grid temperature range = Teff = 5000–15000 K (step 500 K)
    Grid does not cover cool M-dwarfs or very hot stars, both possible in RAVE; detection fraction for such systems is unquantified (Section 3.1).
  • Rotation grid sampling = (Vsini)=1 and U(1,380) km/s (two values)
    The coarse two-value rotation grid may poorly resolve true vsini, weakening the (Vsini)-based selection (Section 3.1).
  • Selection thresholds = fimp≥0.1 / >0.20; χ²_binary<10; χ²_ratio>1.2; S/N>15; σRV1+σRV2<20 km/s; vsini-based inequality
    These hand-set cuts (from Kovalev et al. 2022, 2024b) determine the catalog size; they are applied without re-calibration to RAVE (Table 1).
axioms (5)
  • domain assumption Binary composite model (Eq. 2–3) with Planck-scaled flux ratio accurately represents unresolved SB2 spectra in RAVE's wavelength range
    If line-blending physics, continuum mismatch, or non-Planckian flux ratios deviate, the improvement factor may mis-rank single vs binary fits.
  • domain assumption Forbidden-region criterion: (Vsini)0 < (Vsini)1+(Vsini)2 identifies SB2s
    Assumes single-star fits must over-broaden to mimic the blended profile; used as selection route 1 (Section 3.2).
  • domain assumption Metallicity is equal for both binary components
    The fit forces [Fe/H]1=[Fe/H]2 (Section 3.1); real SB2s with differing metallicities (rare but possible, e.g., post-mass-transfer systems) are mis-modeled.
  • domain assumption RAVE DR6 rest-frame correction (HRV) is accurate for most spectra
    RVs are measured from rest-frame-corrected spectra; the author also identifies a subset with wavelength-calibration distortions (Section 3.3.5), which can create false SB2-like fits.
  • domain assumption Visual inspection by the author is a reliable final classifier
    The final counts are set after inspecting ~12,000 fit plots (Table 1); this is subjective and not reproducible from the text alone.

reviewed 2026-08-02 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Search for double-line spectroscopic binaries in RAVE survey." pith.science (2026). https://pith.science/paper/5H5UFM56

@misc{pith2026260714794,
  author       = {Pith},
  title        = {Pith review of: Search for double-line spectroscopic binaries in RAVE survey},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5H5UFM56}},
  note         = {Machine review of arXiv:2607.14794}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

I present a study for double-line spectroscopic binaries content in spectra from the RAVE spectra, using composite spectral model. Three complementing selections found 2813 composite spectra, belonging to 2605 unique targets. Also these results were compared with automatic classification based on t-SNE map, which shows good agreement. Additionally I identified several spectra affected by problems with wavelength calibration.

Figures

Figures reproduced from arXiv: 2607.14794 by Mikhail Kovalev (YNAO).

Figure 1
Figure 1. Figure 1 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Different selections of SB2 candidates: top left - Matijeviˇc et al. (2010), top right - automatic classification with P=’b’, bottom left - (V sin i) selection, bottom right - primary selection. . Diagonal solid lines show functions (V sin i)1 + (V sin i)2 = (V sin i)0 and (V sin i)1 + (V sin i)2 = 2(V sin i)0 [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: |∆RV| comparison from this work and values from Matijeviˇc et al. (2010). there which is not surprising as this method is similar to t-SNE), although some of them located even at the bottom of the map, so pure automatic classification can miss them. 3.3.3. SX Oph and HD 20784 Detached eclipsing binary SX Oph has 9 spectra in RAVE DR6 under id=J161755.5-063952. I show my RV and light curve (LC) from ASAS-SN… view at source ↗
Figure 4
Figure 4. Figure 4: t-SNE map and different selections of SB2 candidates: top left - Matijeviˇc et al. (2010), top right - automatic classification with P=’b’, bottom left - (V sin i) selection, bottom right - primary selection. while secondary recedes and inverted behavior is seen after shallower eclipse). One measurement taken at the end of primary eclipse was poorly fitted by binary model: RV1 RV0, while RV2 is significant… view at source ↗
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: RAVE spectra and fits for two spectra with wave￾length calibration problems: normal star J051136.3-293551 (top panel) and SB2 candidate J134949.3-621355 (bottom panel). nary model fits such problematic spectra a bit better than single-star model, so they get selected as SB2 can￾didates. Fortunately these spectra are clustered in t￾SNE map at small area around coordinates 97,83 (238 spectra in total) and ca… view at source ↗
Figure 8
Figure 8. Figure 8: Orbital solutions from literature and RV measure￾ments for J161644.2-032302 (top panel), J060757.2-042424 (middle panel) and J132855.2-281836 (bottom panel). and (V sin i)0 shows variability with time. It is similar to Kovalev et al. (2024c), where twin system was discov￾ered in LAMOST-MRS spectra. In this case there will be more detections. I present a catalog with results on RAVE DR6 spectra analysis in … view at source ↗

discussion (0)

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

Reference graph

Works this paper leans on

23 extracted references · 10 canonical work pages

  1. [1]

    K., et al

    Birko, D., Zwitter, T., Grebel, E. K., et al. 2019, AJ, 158, 155, doi: 10.3847/1538-3881/ab3cc1

    work page doi:10.3847/1538-3881/ab3cc1 2019

  2. [2]

    2018, A&A, 616, A5, doi: 10.1051/0004-6361/201832763 De Silva, G

    Cropper, M., Katz, D., Sartoretti, P., et al. 2018, A&A, 616, A5, doi: 10.1051/0004-6361/201832763 De Silva, G. M., Freeman, K. C., Bland-Hawthorn, J., et al. 2015, MNRAS, 449, 2604, doi: 10.1093/mnras/stv327

    doi:10.1051/0004-6361/201832763 2018

  3. [3]

    2018a, MNRAS, 473, 5043, doi: 10.1093/mnras/stx2758

    El-Badry, K., Rix, H.-W., Ting, Y.-S., et al. 2018a, MNRAS, 473, 5043, doi: 10.1093/mnras/stx2758

    doi:10.1093/mnras/stx2758

  4. [4]

    2018b, MNRAS, 476, 528, doi: 10.1093/mnras/sty240 Gaia Collaboration, Brown, A

    El-Badry, K., Ting, Y.-S., Rix, H.-W., et al. 2018b, MNRAS, 476, 528, doi: 10.1093/mnras/sty240 Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1, doi: 10.1051/0004-6361/201833051 Gaia Collaboration, Vallenari, A., Brown, A. G. A., et al. 2022, arXiv e-prints, arXiv:2208.00211. https://arxiv.org/abs/2208.00211

    Pith/arXiv arXiv 2018

  5. [5]

    2012, The Messenger, 147, 25

    Gilmore, G., Randich, S., Asplund, M., et al. 2012, The Messenger, 147, 25

    2012

  6. [6]

    2025, ApJS, 278, 46, doi: 10.3847/1538-4365/adced1

    Guo, S., Kovalev, M., Li, J., et al. 2025, ApJS, 278, 46, doi: 10.3847/1538-4365/adced1

    doi:10.3847/1538-4365/adced1 2025

  7. [7]

    2019, Astronomische Nachrichten, 340, 386, doi: 10.1002/asna.201913496

    Jack, D. 2019, Astronomische Nachrichten, 340, 386, doi: 10.1002/asna.201913496

    work page doi:10.1002/asna.201913496 2019

  8. [8]

    S., Stanek, K

    Jayasinghe, T., Kochanek, C. S., Stanek, K. Z., et al. 2018, MNRAS, 477, 3145, doi: 10.1093/mnras/sty838

    doi:10.1093/mnras/sty838 2018

  9. [9]

    2024a, MNRAS, 527, 9595, doi: 10.1093/mnras/stad3833

    Kovalev, M., Ahmed, M., & Asa’d, R. 2024a, MNRAS, 527, 9595, doi: 10.1093/mnras/stad3833

    work page doi:10.1093/mnras/stad3833

  10. [10]

    2022, MNRAS, 517, 356, doi: 10.1093/mnras/stac2513

    Kovalev, M., Chen, X., & Han, Z. 2022, MNRAS, 517, 356, doi: 10.1093/mnras/stac2513

    doi:10.1093/mnras/stac2513 2022

  11. [11]

    2022, MNRAS, 510, 1515, doi: 10.1093/mnras/stab3365

    Kovalev, M., & Straumit, I. 2022, MNRAS, 510, 1515, doi: 10.1093/mnras/stab3365

    work page doi:10.1093/mnras/stab3365 2022

  12. [12]

    2024b, MNRAS, 527, 521, doi: 10.1093/mnras/stad3222

    Kovalev, M., Zhou, Z., Chen, X., & Han, Z. 2024b, MNRAS, 527, 521, doi: 10.1093/mnras/stad3222

    work page doi:10.1093/mnras/stad3222

  13. [13]

    Y., Chen, X., & Han, Z

    Kovalev, M. Y., Chen, X., & Han, Z. 2024c, Research Notes of the American Astronomical Society, 8, 175, doi: 10.3847/2515-5172/ad5f2f

    work page doi:10.3847/2515-5172/ad5f2f

  14. [14]

    Y., Kniazev, A

    Kovalev, M. Y., Kniazev, A. Y., & Malkov, O. Y. 2026, Galaxies, 14, doi: 10.3390/galaxies14020027

    work page doi:10.3390/galaxies14020027 2026

  15. [15]

    2020, arXiv e-prints, arXiv:2005.07210

    Liu, C., Fu, J., Shi, J., et al. 2020, arXiv e-prints, arXiv:2005.07210. https://arxiv.org/abs/2005.07210 Matijeviˇ c, G., Zwitter, T., Munari, U., et al. 2010, AJ, 140, 184, doi: 10.1088/0004-6256/140/1/184 Matijeviˇ c, G., Zwitter, T., Bienaym´ e, O., et al. 2011, AJ, 141, 200, doi: 10.1088/0004-6256/141/6/200 —. 2012, ApJS, 200, 14, doi: 10.1088/0067-0...

    Pith/arXiv arXiv 2020

  16. [16]

    A., Batten, A

    Pourbaix, D., Tokovinin, A. A., Batten, A. H., et al. 2004, A&A, 424, 727, doi: 10.1051/0004-6361:20041213

    doi:10.1051/0004-6361:20041213 2004

  17. [17]

    M., Jayasinghe, T., Stanek, K

    Rowan, D. M., Jayasinghe, T., Stanek, K. Z., et al. 2022, MNRAS, 517, 2190, doi: 10.1093/mnras/stac2520

    doi:10.1093/mnras/stac2520 2022

  18. [18]

    2018, Research Notes of the American Astronomical Society, 2, 194, doi: 10.3847/2515-5172/aaead0

    Munari, U. 2018, Research Notes of the American Astronomical Society, 2, 194, doi: 10.3847/2515-5172/aaead0

    work page doi:10.3847/2515-5172/aaead0 2018

  19. [19]

    2020, AJ, 160, 82, doi: 10.3847/1538-3881/ab9ab9

    Steinmetz, M., Matijeviˇ c, G., Enke, H., et al. 2020, AJ, 160, 82, doi: 10.3847/1538-3881/ab9ab9

    doi:10.3847/1538-3881/ab9ab9 2020

  20. [20]

    2018, A&A, 620, A71, doi: 10.1051/0004-6361/201833228

    Sysoliatina, K., Just, A., Koutsouridou, I., et al. 2018, A&A, 620, A71, doi: 10.1051/0004-6361/201833228

    work page doi:10.1051/0004-6361/201833228 2018

  21. [21]

    2026, arXiv e-prints, arXiv:2601.21125, doi: 10.48550/arXiv.2601.21125

    Tanikawa, A., Tajitsu, A., Honda, S., et al. 2026, arXiv e-prints, arXiv:2601.21125, doi: 10.48550/arXiv.2601.21125

    Pith/arXiv arXiv doi:10.48550/arxiv.2601.21125 2026

  22. [22]

    Taylor, M. B. 2005, in Astronomical Society of the Pacific Conference Series, Vol. 347, Astronomical Data Analysis Software and Systems XIV, ed. P. Shopbell, M. Britton, & R. Ebert, 29

    2005

  23. [23]

    2017, ApJS, 228, 24, doi: 10.3847/1538-4365/228/2/24

    Traven, G., Matijeviˇ c, G., Zwitter, T., et al. 2017, ApJS, 228, 24, doi: 10.3847/1538-4365/228/2/24

    work page doi:10.3847/1538-4365/228/2/24 2017

This paper was first reviewed by deepseek-v4-flash on August 2, 2026.