Matched filtering of gravitational waves from inspiraling compact binaries: Computational cost and template placement

2) ((1) Caltech; (2) Cardiff); Benjamin J. Owen (1); B. S. Sathyaprakash (1

REVIEW

Matched filtering of gravitational waves from inspiraling compact binaries: Computational cost and template placement

Benjamin J. Owen (1), B. S. Sathyaprakash (1, 2) ((1) Caltech, (2) Cardiff)

General Relativity and Quantum Cosmology · Submitted 27 Aug 1998 · 34 references

Pith's verdict: A solid, honest 1998 resource-budget paper: the 2PN template counts and flops figures are the useful contribution, and the soft spots are real but don't move the headline numbers.

T0 · AI review

How this review was made: An AI referee (selfhosted/deepseek-v4.1-flash) read the paper on October 29, 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 →

· 2 major objections · 7 minor · Ask about this paper · The compiled record · Read the full review →

Challenge this review Comment (0) Read on arXiv ↗

Cite ▾

arXiv IDgr-qc/9808076 (version 1)

Pith IDY6JHQP55, Pith's permanent record of this paper and its review

Citation2) ((1) Caltech, (2) Cardiff), Benjamin J. Owen (1), B. S. Sathyaprakash (1. "Matched filtering of gravitational waves from inspiraling compact binaries: Computational cost and template placement." arXiv:gr-qc/9808076 (1998). https://pith.science/paper/Y6JHQP55

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

This is the paper people cite when they need a number for how much compute an inspiral search eats. That's the whole point of it, and it does that job well. The update over Owen (1996) is genuine: 2PN restricted waveforms instead of 1PN, six detector noise curves instead of one, and a properly justified sampling-rate criterion (Eq. 3.6, credited to Flanagan) rather than a hand-wave. The chain from Eq. (3.7) through the metric to the proper volume to flops is written out and the arithmetic checks — I reproduced LIGO I from Table II and Eq. (2.18) and got Table III's 9.9e10. The noise fits are printed in Table I, so even the curves that came in by private communication are reproducible as stated. Sec. IV's greedy placement along the equal-mass curve is a sensible scheme, and the authors are upfront that it's a sketch rather than a validated algorithm. They also flag their own scope limits in Sec. V: precessing binaries untouched, the minimal-match-to-event-rate relation crude. The classifier says no limitations were stated; that's wrong, the discussion section states several. On the reader's main worry — the quadratic match approximation degrading for massive binaries at 2PN, per the authors' own Fig. 1 caption — I think it's the right thing to notice but it doesn't bite. The proper volume scales as m_min^{-8/3}, so the count is set by the low-mass corner where Fig. 1 shows good agreement, and Sec. II B says the approximation generally underestimates the match, i.e. inflates N. The residual gap is directional: the caption doesn't say which way the massive-binary discrepancy goes, and only an overestimate of the match would make the count too small. Bounded, not load-bearing. The one criticism I'd actually press is that the intro argues the old factor-of-3 accuracy is no longer adequate, then attaches no uncertainty band to any headline number. That's a fair hit, and it's the reason these figures get read as order-of-magnitude planning numbers rather than budgets. Who should read it: anyone building or benchmarking an inspiral search pipeline, and anyone writing a hierarchical-search proposal that needs the one-step baseline. Worth a serious referee — the numbers are checkable and the method is stated.

Referee report

Significance

If the numbers hold, this is a useful and durable benchmark: it tells the community what a brute-force inspiral search costs before any hierarchical or multi-detector refinement, and it is framed explicitly as the baseline against which such refinements are to be judged. Several features deserve credit. All six noise spectral densities are given as closed-form fits in Table I with S0, f0, fs, and fu tabulated, so the inputs are visible and the calculation is reproducible rather than opaque. The arithmetic is checkable and checks out: for LIGO I, N = 4.8e5, fu = 1300 Hz, and log2 F = 21 give P = 4.8e5 x 1300 x 158 = 9.9e10 flops and S = N F = 1.0e12, both matching Table III, and the same holds for VIRGO and LIGO III. The scaling laws of Eqs. (3.11)-(3.14) are falsifiable predictions that later work can test, and the mmin^{-8/3} and mmin^{-13/3} scalings are the kind of statement that remains useful even if the prefactors are revised. The paper is also candid about what it does not do: precessing binaries, spin, and eccentricity are explicitly deferred, and the authors state that low-mass searches will likely need on-the-fly template generation rather than storage.

Major objections (2)

[Sec. III C, Tables II-IV, and Fig. 1] The paper's stated motivation in Sec. I is that 'the factor of 3 accuracy attempted in Ref. [23] is no longer adequate,' yet Tables II-IV quote three significant figures for N, P, and S with no uncertainty attached anywhere. The one approximation the manuscript itself flags as failing somewhere is the quadratic match model of Eq. (2.9): the Fig. 1 caption states that at 2PN order the agreement with the numerically computed match 'is quite good for all but waveforms of massive binaries.' Because the proper volume scales as mmin^{-8/3} (Eqs. 3.11-3.12), the low-mass corner dominates the integral and the headline numbers are probably safe, but the paper should say this explicitly rather than leave the reader to reconstruct it. It should also state the direction of the Fig. 1 discrepancy for massive binaries: Sec. II B argues the quadratic form 'generally underestimates the true match' and so errs conservatively, but if the massive-binary deviation goes the other way, the template count in that corner is too small. A sentence quantifying the fractional contribution of the M > 20 solar-mass region to the integral in Eq. (3.8), plus the sign of the discrepancy, would close this gap cheaply.
[Table I and Sec. III B] The six noise spectral densities in Table I are the sole physical input that sets N, P, and S, and all three scale with f0 and fs through Eqs. (3.12)-(3.14). Yet no statement is given of how faithfully the analytic fits reproduce the underlying reference spectra. Two of the six fits are attributed to private communications (Thorne for LIGO I, Hughes for LIGO III) and one to a preprint listed as 'to appear' (Ref. [32], gr-qc/9804044), which a reader cannot presently check. A short table of maximum fractional residuals over the relevant band, or at minimum a sentence per detector stating the fit accuracy, would let the reader judge whether the quoted three significant figures are meaningful or whether the budget carries an unstated systematic of tens of percent. This is not a request to redo the detector modeling; it is a request to state the fidelity of an input on which the entire result rests.

Minor objections (7)

[Table II] The template counts are non-monotonic in post-Newtonian order: for every detector N(1PN) < N(1.5PN) > N(2PN), e.g. LIGO I gives 2.3e5, 5.6e5, 4.8e5. The 1.5PN-to-2PN change is about 14%, which supports the claim that the sequence is converging, but the paper never remarks on this. One sentence noting the non-monotonicity and its small size at the last step would strengthen the case that the 2PN count is adequate.
[Abstract and Sec. I] The abstract states the new numbers are 'several times higher than estimated in previous work by Owen, in part because of the improved family of templates and in part because we use more realistic (higher) sampling rates.' Table II supports the template-family part of this decomposition (1PN versus 2PN columns), but the sampling-rate contribution is never separated out. Since the comparison to Ref. [23] is one of the paper's stated contributions, a sentence giving the two effects separately would make the claim checkable.
[Sec. II C, Eq. (2.10), and Sec. IV] Equation (2.10) is stated to underestimate N because of template spillover, and Sec. IV says the total number 'must be modestly greater' than the analytic estimate, but no bound is given. Similarly, the hexagonal-lattice reduction of about 20% is said to be partly offset by the details of constructing a lattice, again without quantification. Since the paper's purpose is to supply a reliable baseline, a rough bound on the net spillover-plus-packing correction (even 'within 30%') would be more useful than the qualitative statements.
[Sec. III B, Eqs. (3.5)-(3.6)] The high-frequency cutoff is set to twice the right-hand side of Eq. (3.6) so that the sampling loss is one quarter of the template loss, and the authors describe this as 'a first cut.' A sentence on how P varies with fu, or on what the optimal choice would be, would help a reader who wants to reuse this criterion for a different detector.
[Sec. IV and Fig. 4] The placement algorithm is described as a sketch, and Fig. 4 gives lattice parameters (dx0 = 46.5 ms, dx1 = 4.45 ms, theta = 0.35 rad) at a single point, with the note that they 'vary slightly from these values at other locations.' It would help to state how much they vary, since the algorithm's practical value depends on whether a single local diagonalization suffices over the whole region or whether it must be redone at each step.
[References] The TAMA noise spectrum is cited to Ref. [31], the grasp software package, rather than to a primary detector paper; if a published TAMA sensitivity reference exists, it should be cited alongside or instead. Reference [32] is listed as 'to appear' and should be updated to its published form if available.
[Text and typography] Several typographical errors should be corrected: 'analagous' (Sec. I), 'refereed to as' for 'referred to as' (Sec. II A), 'interferometeric' (Sec. V), 'hiearchical' (Sec. V), and 'Durandhar' for 'Dhurandhar' (Sec. I). The axis labels and units of Figs. 2 and 4 are not legible in the present rendering and should be checked.

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

The paper

The paper claims

A matched-filter search for inspiraling binaries is priced out in templates, flops, and storage, and the bill is large enough to shape how the searches get built.

The reading

The paper asks a practical question that sits between gravitational-wave theory and computer engineering: if you want to catch the chirp of two compact objects spiraling together, how many waveform templates must you run against the data, and what will that cost in computing power and memory? The authors answer by turning the space of possible binary masses into a curved geometry, where distance measures how much signal-to-noise you lose by using the wrong template, and then computing the proper volume of the region that must be covered. Their headline number is that initial LIGO would need roughly $1.0^{11}$ floating-point operations per second just to keep pace with the data stream, several times the earlier estimate, with advanced LIGO at $7.8^{11}$ and VIRGO at $4.8^{12}$. What makes the number matter is that it is offered as a benchmark: any cleverer strategy, such as a hierarchical search that first passes the data through a coarse grid, has to be measured against this baseline. The paper also sketches how to lay the templates down along the equal-mass line and stack rows outward to cover the mass plane.

What carries the argument

The load-bearing device is a metric on the space of intrinsic binary parameters, built from the match function $M( heta_1, heta_2)$ — the fraction of optimal signal-to-noise recovered by filtering with a template whose parameters differ from the signal's, after maximizing over the unknown coalescence time and phase. Expanding the match about a coincident point gives $M simeq 1 - g_{ij} Delta heta^i Delta heta^j$, so the metric measures how quickly detectability degrades as you move away from the true parameters. The number of templates is then the proper volume of the region of interest divided by the proper volume per lattice cell, with the cell size set by the minimal match and the packing scheme (a rectangle in two dimensions, or a hexagon for a 20% saving). To keep the metric components from varying wildly, the authors use chirp-time coordinates rather than the two masses, and they project out the coalescence time before integrating. A second piece of machinery is the sampling-rate condition: the high-frequency cutoff must be high enough that discretization in time costs less signal-to-noise than discretization in template parameters, which is why their sampling rates, and hence their flop counts, are higher than in earlier work.

Extended reading notes

Core claim

The central claim is a resource budget for a one-step matched-filter search using restricted second post-Newtonian waveforms for nonspinning binaries in circular orbits. Covering all binaries with component masses above $0.2M_*$ while losing no more than 10% of events to coarse template spacing requires about $4.8^{5}$ templates for initial LIGO, $6.9^{6}$ for advanced LIGO, and $9.3^{6}$ for VIRGO on a rectangular lattice at 97% minimal match. Translated through a standard operations count for Fourier-domain filtering, that becomes $9.9^{10}$ flops for initial LIGO, $7.8^{11}$ for advanced LIGO, and $4.8^{12}$ for VIRGO, with stored templates running from $1.5^{11}$ real numbers for TAMA to $6.2^{14}$ for VIRGO. The authors also derive scaling laws: computational power goes roughly as $m_{min}^{-8/3}$ and storage as $m_{min}^{-13/3}$, so pushing the mass threshold down is expensive fast. One pointed corollary they draw is that VIRGO's very broad, flat noise spectrum cannot be fully exploited by this kind of search, because the low-frequency reach multiplies the template count beyond what the available computing can service.

Load-bearing premise

The entire template count rests on treating the match between neighbouring templates as a quadratic bowl out to a 3% mismatch, and the authors' own Fig. 1 caption concedes that at second post-Newtonian order this agreement holds for all but waveforms of massive binaries, which is precisely the corner of parameter space the proper-volume integral includes.

Editorial extensions

If this is right

Any proposed shortcut — hierarchical two-stage searches, template interpolation, reduced-order waveform generation — now has a concrete number to beat rather than a vague sense that the naive search is costly.
Detector designers get a warning that a broader or flatter noise spectrum buys sensitivity at a superlinear price in analysis cost, since template count scales as $f_0^{-8/3}$ and power as $f_0^{-5/3}$.
Searches extending down to subsolar component masses, such as low-mass black hole dark matter candidates, are effectively forced to generate templates on the fly rather than store them, because storage scales as $m_{min}^{-13/3}$.
The template-placement algorithm, which walks along the equal-mass curve and stacks rows outward in locally diagonalized coordinates, gives an implementable recipe rather than only a count, including the expected boundary spill-over.
The 97% minimal match is tied to a 10% loss of event rate through the cube of the fitting factor, giving a single tunable dial connecting search thoroughness to computational budget.

Reading between the lines

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

The budget is a moving target in a useful sense: because power scales as $m_{min}^{-8/3}$, raising the mass floor from $0.2M_*$ to $1M_*$ cuts the requirement by roughly two orders of magnitude, which suggests the mass threshold is the single most consequential design choice in an inspiral search and deserves explicit justification rather than convention.
The same geometric machinery should transfer almost unchanged to any waveform family with a smooth parameter dependence, so the metric-plus-proper-volume method is plausibly reusable for eccentricity, spin-aligned, or precessing searches once those waveforms are cheap enough to evaluate.
The paper's own numbers imply that the decisive engineering question for a real pipeline is not the flop count but the ratio of data-segment length to filter length, since the operations count carries a $log_2 F$ term that the paper fixes rather than optimizes.
If a hierarchical search really delivers the factor-of-25 reduction attributed to it in the literature, then the naive budget in this paper is comfortably affordable at initial-LIGO sensitivity, which reframes the numbers here as an upper bound on what the problem costs rather than a barrier.

Original abstract

We estimate the number of templates, computational power, and storage required for a one-step matched filtering search for gravitational waves from inspiraling compact binaries. These estimates should serve as benchmarks for the evaluation of more sophisticated strategies such as hierarchical searches. We use waveform templates based on the second post-Newtonian approximation for binaries composed of nonspinning compact bodies in circular orbits. We present estimates for six noise curves: LIGO (three configurations), VIRGO, GEO600, and TAMA. To search for binaries with components more massive than 0.2M_o while losing no more than 10% of events due to coarseness of template spacing, initial LIGO will require about 1*10^11 flops (floating point operations per second) for data analysis to keep up with data acquisition. This is several times higher than estimated in previous work by Owen, in part because of the improved family of templates and in part because we use more realistic (higher) sampling rates. Enhanced LIGO, GEO600, and TAMA will require computational power similar to initial LIGO. Advanced LIGO will require 8*10^11 flops, and VIRGO will require 5*10^12 flops. If the templates are stored rather than generated as needed, storage requirements range from 1.5*10^11 real numbers for TAMA to 6*10^14 for VIRGO. We also sketch and discuss an algorithm for placing the templates in the parameter space.

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-1.13-20260917 via OpenRouter, question bank abstract-p1-20260923, 2026-09-23: 1 request, 7,632 input tokens, $0.0003 reported.

What the abstract itself states, as probabilities: claims first 0.04 · states limitation 0.49 · human participants 0.01 · controlled comparison 0.14 · randomized assignment 0.01 · states quantitative result 0.99 · compares with prior results 0.97 · mentions code or data availability 0.02.

The abstract, sentence by sentence

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

We estimate the number of templates, computational power, and storage required for a one-step matched filtering search for gravitational waves from inspiraling compact binaries. These estimates should serve as benchmarks for the evaluation of more sophisticated strategies such as hierarchical searches. We use waveform templates based on the second post-Newtonian approximation for binaries composed of nonspinning compact bodies in circular orbits. We present estimates for six noise curves: LIGO (three configurations), VIRGO, GEO600, and TAMA. To search for binaries with components more massive than 0.2M_o while losing no more than 10% of events due to coarseness of template spacing, initial LIGO will require about 1*10^11 flops (floating point operations per second) for data analysis to keep up with data acquisition. This is several times higher than estimated in previous work by Owen, in part because of the improved family of templates and in part because we use more realistic (higher) sampling rates. Enhanced LIGO, GEO600, and TAMA will require computational power similar to initial LIGO. Advanced LIGO will require 8*10^11 flops, and VIRGO will require 5*10^12 flops. If the templates are stored rather than generated as needed, storage requirements range from 1.5*10^11 real numbers for TAMA to 6*10^14 for VIRGO. We also sketch and discuss an algorithm for placing the templates in the parameter space.

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	We estimate the number of templates, computational power, and storage required for a one-step matched filtering search for gravitational waves from inspiraling compact binaries.	contribution 1.00	0.92	0.08	asserted 0.97
s1	These estimates should serve as benchmarks for the evaluation of more sophisticated strategies such as hierarchical searches.	interpretation 0.54	0.17	0.08	asserted 0.49
s2	We use waveform templates based on the second post-Newtonian approximation for binaries composed of nonspinning compact bodies in circular orbits.	method 1.00	0.71	0.03	asserted 1.00
s3	We present estimates for six noise curves: LIGO (three configurations), VIRGO, GEO600, and TAMA.	contribution 0.51	0.87	0.04	asserted 1.00
s4	To search for binaries with components more massive than 0.2M_o while losing no more than 10% of events due to coarseness of template spacing, initial LIGO will require about 1*10^11 flops (floating point operations per second) for data analysis to keep up with data acquisition.	result 1.00	0.91	0.28	asserted 0.83
s5	This is several times higher than estimated in previous work by Owen, in part because of the improved family of templates and in part because we use more realistic (higher) sampling rates.	result 0.51	0.65	0.07	asserted 0.97
s6	Enhanced LIGO, GEO600, and TAMA will require computational power similar to initial LIGO.	result 1.00	0.85	0.06	asserted 1.00
s7	Advanced LIGO will require 810^11 flops, and VIRGO will require 510^12 flops.	result 1.00	0.90	0.18	asserted 1.00
s8	If the templates are stored rather than generated as needed, storage requirements range from 1.510^11 real numbers for TAMA to 610^14 for VIRGO.	result 1.00	0.90	0.07	asserted 0.88
s9	We also sketch and discuss an algorithm for placing the templates in the parameter space.	contribution 0.57	0.89	0.03	asserted 0.89

What kind of paper the classifier reports this is: simulation study 0.57 · evidence from simulation or computation 0.95.

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.** FIG. 1. Comparison of the match to the quadratic approximati

**Figure 2.** FIG. 2. The parameter space of search in terms of (a) standard

**Figure 3.** FIG. 3. Illustration of optimal translation of a template al

**Figure 4.** FIG. 4. Choice of templates in the space of chirp times for ini

Figures reproduced from arXiv:gr-qc/9808076 by the authors.

References

Works cited (34)

[23]

B. J. Owen, Phys. Rev. D 53, 6749 (1996)

View on Pith 1996
[32]

B. J. Owen, L. Lindblom, C. Cutler, B. F. Schutz, A. Vecch io, and N. Andersson, to appear in Phys. Rev. D (gr-qc/980404 4)

View on Pith
[1]

A. Abramovici et al. , Science 256, 325 (1992). 14

View on Pith 1992
[2]

B. Caron et al. , Class. Quantum Grav. 14, 1461 (1997)

View on Pith 1997
[3]

H. L¨ ucket al. , Class. Quantum Grav. 14, 1471 (1997)

View on Pith 1997
[4]

K. Tsubono, in Gravitational Wave Experiments (World Scientiﬁc, Singapore, 1995), p. 112

View on Pith 1995
[5]

R. Narayan, T. Piran, and A. Shemi, Astrophys. J. 379, L17 (1991); E. S. Phinney, Astrophys. J. 380, L17 (1991). In the intervening years astronomers have estimated event rates b oth lower (based on statistics of the few known binary pulsar s) and higher (based on theoretical evolutions of progenitor p opulations), but their estimates have begun to conver...

View on Pith 1991
[6]

S. F. P. Zwart and H. N. Spreeuw, Astron. Astrophys. 312, 670 (1996)

View on Pith 1996
[7]

E. P. J. van den Heuvel and D. R. Lorimer, Mon. Not. R. Astro n. Soc. 283, L67 (1996); I. H. Stairs et al., astro-ph/9712296; V. M. Lipunov, K. A. Postnov and M. E. Prokhorov, Mon. Not. R. A stron. Soc. 288, 245 (1997)

View on Pith Pith review arXiv 1996
[8]

B. Barish et al. , LIGO Advanced Research and Development Proposal, Caltech /MIT, 1996 (unpublished)

View on Pith 1996
[9]

L. Blanchet, T. Damour, B. R. Iyer, C. M. Will, and A. G. Wis eman, Phys. Rev. Lett. 74, 3515 (1995)

View on Pith 1995
[10]

H. Tagoshi, M. Shibata, T. Tanaka, and M. Sasaki, Phys. R ev. D 54, 1439 (1996)

View on Pith 1996
[11]

Some physicists are in the habit of calling this optimal ﬁltering method simply “Wiener ﬁltering.” However, in the l arger signal processing community, “Wiener ﬁltering” refers to a n altogether diﬀerent technique. In order to avoid confusio n we call this method “matched ﬁltering.”

View on Pith
[12]

C. W. Helstrom, Statistical Theory of Signal Detection , 2nd edition (Pergamon Press, London, 1968)

View on Pith 1968
[13]

T. A. Apostolatos, Phys. Rev. D 52, 605 (1995)

View on Pith 1995
[14]

S. Droz and E. Poisson, Phys. Rev. D 56, 4449 (1997)

View on Pith 1997
[15]

T. Damour, B. R. Iyer and B. S. Sathyaprakash, Phys. Rev. D 57, 885 (1998)

View on Pith 1998
[16]

R. Balasubramanian and S. V. Dhurandhar, Phys. Rev. D 50, 6080 (1994)

View on Pith 1994
[17]

A. Kr´ olak, K. D. Kokkotas, and G. Sch¨ afer, Phys. Rev. D 52, 2089 (1995)

View on Pith 1995
[18]

T. A. Apostolatos, Phys. Rev. D 54, 2421 (1996)

View on Pith 1996
[19]

B. S. Sathyaprakash, Phys. Rev. D 50, R7111 (1994)

View on Pith 1994
[20]

B. S. Sathyaprakash and S. V. Dhurandhar, Phys. Rev. D 44, 3819 (1991)

View on Pith 1991
[21]

B. S. Sathyaprakash and S. V. Dhurandhar, J. Comput. Phy s. 109, 215 (1993)

View on Pith 1993
[22]

S. V. Dhurandhar and B. S. Sathyaprakash, Phys. Rev. D 49, 1707 (1994)

View on Pith 1994
[24]

B. F. Schutz, in The Detection of Gravitational Radiation, edited by D. Blair (Cambridge University Press, Cambridge, England, 1991), p. 406

View on Pith 1991
[25]

S. D. Mohanty and S. V. Dhurandhar, Phys. Rev. D 54, 7108 (1996)

View on Pith 1996
[26]

S. D. Mohanty, Phys. Rev. D 57, 630 (1998)

View on Pith 1998
[27]

C. Cutler and ´E. ´E. Flanagan, Phys. Rev. D 49, 2658 (1994)

View on Pith 1994
[28]

R. Balasubramanian, B. S. Sathyaprakash and S. V. Dhura ndhar, Phys. Rev. D 53, 3033 (1996); 54, 1869(E) (1996)

View on Pith 1996
[29]

E. Poisson and C. M. Will, Phys. Rev. D 52, 848 (1995)

View on Pith 1995
[30]

P. C. Peters, Phys. Rev. 136, B1224 (1964)

View on Pith 1964
[31]

B. Allen, grasp software package, available at http://www.ligo.caltech. edu/LIGO web/lsc/lsc.html

View on Pith
[33]

T. Nakamura, M. Sasaki, T. Tanaka, and K. S. Thorne, Astr ophys. J. 487, L139 (1997)

View on Pith 1997
[34]

J. H. Taylor, Rev. Mod. Phys. 66, 711 (1994). 15

View on Pith 1994

Forward citations

Cited by 21 Pith papers

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

Multiband parameter estimation with phase coherence and extrinsic marginalization: Extracting more information from low-SNR CBC signals in LISA data
gr-qc 2025-06 conditional novelty 8.0 of 10

A coherent multiband Bayesian parameter estimation method with extrinsic-parameter marginalization extracts useful information from LISA observations of stellar-mass binary black holes down to LISA SNR 3, nearly doubl...
Accelerated Time-domain Analysis for Gravitational Wave Astronomy
gr-qc 2026-03 unverdicted novelty 7.0 of 10

Presents a practical fully time-domain end-to-end likelihood for gravitational-wave inference with structured linear algebra and GPU acceleration.
A geometric template bank for the detection of spinning low-mass compact binaries with moderate orbital eccentricity
gr-qc 2024-12 conditional novelty 7.0 of 10

A geometric template bank including eccentricity recovers moderately eccentric BNS/NSBH signals with less than 6% loss, while quasi-circular banks miss up to 40%.
The Bondi--Sachs gauge, BMS frames, and memory in black hole perturbation theory
gr-qc 2026-06 unverdicted novelty 6.0 of 10

Introduces a gauge transformation framework for BMS frames in multiscale black hole perturbation theory on Kerr that incorporates memory effects and avoids infrared divergences.
Improving low-latency multi-messenger follow-up of neutron star-black hole mergers with mode-by-mode filtering
gr-qc 2026-06 unverdicted novelty 6.0 of 10

Mode-by-mode filtering of higher-order modes enables low-latency marginalization over mode information in NSBH gravitational-wave signals, tightening constraints on distance, inclination, and secondary mass.

Show 16 more citing papers

Impact of numerical-relativity waveform calibration on parametrized post-Einsteinian tests
gr-qc 2026-03 accept novelty 6.0 of 10

NR late-inspiral calibration systematics in IMRPhenomD produce false ppE GR violations at O5 SNRs ≳60; an uncertainty-aware baseline restores consistency with GR up to SNR 330.
Identifying Microlensing by Compact Dark Matter through Diffraction Patterns in Gravitational Waves with Machine Learning
astro-ph.IM 2025-09 conditional novelty 6.0 of 10

A wavelet-convolution neural network distinguishes simulated lensed from unlensed gravitational waves with 92.2% accuracy (AUC 0.965) using wave-optics diffraction patterns.
Template bank for sub solar mass compact binary mergers in the fourth observing run of Advanced LIGO, Advanced Virgo, and KAGRA
gr-qc 2024-12 conditional novelty 6.0 of 10

Two template banks for sub-solar-mass compact binary searches in LIGO-Virgo-KAGRA's fourth observing run were generated with a modified manifold algorithm, and simulation tests show sufficient match coverage for masse...
Deep Neural Networks to Enable Real-time Multimessenger Astrophysics
astro-ph.IM 2016-12 conditional novelty 6.0 of 10

A pair of one-dimensional convolutional neural networks detects simulated binary black hole signals at signal-to-noise ratio 10 and estimates component masses to roughly 10 percent, running in microseconds on a GPU.
Searching for Gravitational Waves from Compact Binaries with Precessing Spins
gr-qc 2016-03 conditional novelty 6.0 of 10

A sky-maximized matched-filter statistic enables searches with precessing-spin templates; averaged over the parameter space it is roughly break-even, but it recovers more signal for highly precessing, high-mass-ratio ...
The PyCBC search for gravitational waves from compact binary coalescence
gr-qc 2015-08 conditional novelty 6.0 of 10

PyCBC's matched-filter search pipeline, documented here, found GW150914 and GW151226 and recovers roughly 30% more binary neutron star sensitive volume than the previous LIGO pipeline on the same data.
Error-analysis and comparison to analytical models of numerical waveforms produced by the NRAR Collaboration
gr-qc 2013-07 conditional novelty 6.0 of 10

A cross-code audit of 25 numerical black-hole merger waveforms finds that the effective-one-body and phenomenological template families used in LIGO/Virgo searches recover them with over 99% overlap at total masses of...
Characterizing the Properties and Constitution of Compact Objects in Gravitational-Wave Binaries
gr-qc 2024-11 conditional novelty 5.0 of 10

Tidal heating imprints from black hole horizons can be captured by two effective parameters and modeled through merger, offering a future test to distinguish black holes from exotic horizonless objects.
Gwbench: a novel Fisher information package for gravitational-wave benchmarking
gr-qc 2020-10 accept novelty 5.0 of 10

A new open-source Python package implements Fisher-information benchmarking for gravitational-wave detector networks, validated against numeric-versus-symbolic derivatives and against published measurements of three r...
Waveform systematics in the gravitational-wave inference of tidal parameters and equation of state from binary neutron star signals
gr-qc 2020-09 conditional novelty 5.0 of 10

Current gravitational-wave waveform models for binary neutron stars disagree by more than statistical errors at SNR above about 80, and by more than one sigma at SNR 100 to 200, limiting equation-of-state inference fo...
Multiband Gravitational-Wave Astronomy: Observing binary inspirals with a decihertz detector, B-DECIGO
gr-qc 2018-02 conditional novelty 5.0 of 10

B-DECIGO alone could pin the mass ratio of a GW150914-like binary to better than 0.1 percent and its black-hole spins to better than 10 percent, with multiband follow-up sharpening neutron-star tidal measurements.
GW170104: Observation of a 50-Solar-Mass Binary Black Hole Coalescence at Redshift 0.2
gr-qc 2017-06 accept novelty 5.0 of 10

A third stellar-mass binary black hole merger is detected at redshift 0.18, with component masses near 31 and 19 solar masses, and every general-relativity null test applied to it passes.
Upper limits on the rates of binary neutron star and neutron-star--black-hole mergers from Advanced LIGO's first observing run
astro-ph.HE 2016-07 accept novelty 5.0 of 10

No binary neutron star or neutron star-black hole mergers were detected in Advanced LIGO's first observing run, yielding 90% confidence rate ceilings of 12,600 and 3,600 per cubic gigaparsec per year.
GWTC-2.1: Deep Extended Catalog of Compact Binary Coalescences Observed by LIGO and Virgo During the First Half of the Third Observing Run
gr-qc 2021-08 accept novelty 4.0 of 10

GWTC-2.1 adds eight new high-significance compact binary coalescence events to the prior catalog, extending the observed black hole mass range and including candidates inside the pair-instability mass gap.
GstLAL O4 Online Results Paper
gr-qc 2026-05 unverdicted novelty 3.0 of 10

GstLAL produced low-latency alerts for 250 astrophysically plausible gravitational-wave candidates during O4, providing the first upload for 222 and the sole upload for 75, with 88 percent of significant catalog event...
Physics, Astrophysics and Cosmology with Gravitational Waves
gr-qc 2009-03 accept novelty 2.0 of 10

The paper reviews the physics, sources, detectors, data analysis and scientific payoff of gravitational wave astronomy, forecasting that upgraded ground-based interferometers should reach regular detections around 2014.

How later papers used it

116 later papers cite it

Evidence: substantial. 108 of the citing papers share no author with it, and 24 of those carry a confident label from Pith's citation classifier.

Pith read 11 of the 177 citing sentences, choosing only those whose citation markers point at this paper alone, and wrote the paraphrases from them. The summary, the counts, the evidence level, the links and the quotes come from code; the summary counts the papers listed below. A later paper is listed under a label when Pith's citation classifier gives one of its sentences that label at probability 0.8 or more (0.95 for support), Pith's reading finds that the citing paper itself did it, and a second check confirms the paraphrase against the sentence. This section lists uses only. It does not list criticism or disagreement, so their absence here says nothing about how the paper has been received. Papers that share an author with this one are listed apart, under "By the same authors". A quote is the citing authors' own sentence, shown when their paper's licence is CC BY, CC BY-SA or CC0. JSON

1 later paper by other authors is listed below: 1 used its method or tool. 1 more paper by the same authors is listed: 1 built on it.

Builds on it 1 of 2 papers with this label · 1 by the same authors

By the same authors

Extended this paper's non-spinning inspiral metric to 3.5 PN order in phase.

Parameter space metric for 3.5 post-Newtonian gravitational-waves from compact binary inspirals (2013) (same authors) · 1.00 · paraphrased: the licence does not allow quoting

Uses its method or tool 1 of 25 papers with this label

Computed the metric by evaluating the Fisher information matrix of this paper over the full parameter space.

Effectual gravitational-wave template banks for coalescing compact binaries using a hybrid placement algorithm (2017) · 0.99 · paraphrased: the licence does not allow quoting

How the literature uses it

80 arXiv papers cite this in 127 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 · 66 sentences without a confident label

Builds on it 2 papers

We present and employ a geometric method to construct a bank of eccentric waveforms for matched-filtering spanning the intrinsic parameter space $(m^_1, m^_2, s^_1z, s^_2z, e^_0)$, building on the works of [ref] .

A geometric template bank for the detection of spinning low-mass compact binaries with moderate orbital eccentricity · Geometric Approach · 0.89
Hence in this work we extend the non-spinning inspiral metric of [ref] and [ref] to 3.5 PN order in phase.

Parameter space metric for 3.5 post-Newtonian gravitational-waves from compact binary inspirals · Introduction · 1.00

Uses its method or tool 19 papers

We used the TaylorF2 second-order post-Newtonian hexagonal placement algorithm of Refs. [ref] to construct a template bank of EOBNRv2 waveforms with $$ of $0.97$.

Template banks to search for low-mass binary black holes in advanced gravitational-wave detectors · Conclusions · 1.00
For details of template bank construction, we use Ref. [ref] .

Detectability of Sub-Solar Mass Neutron Stars Through a Template Bank Search · Impact on Search Sensitivity · 0.99
We first generate a template bank for binaries with zero eccentricity and component spin using the standard hexagonal lattice method of template placement [ref] .

Eccentric Binary Neutron Star Search Prospects for Cosmic Explorer · Binary Neutron Star Searches in Cosmic Explorer · 1.00

Compares with it 1 paper

The size of these banks is comparable to one constructed using the second-order post-Newtonian TaylorF2 hexagonal template placement method [ref] , which yields a grid of $522$ and $736$ templates, for a minimal match of $97%$ and $98%$, respectively.

Template Banks for Binary black hole searches with Numerical Relativity waveforms · Constructing a template bank for NR-PN hybrids · 0.98

Cites it as support 1 paper

* Mass lower bound — The lower bound of our mass region is informed by the fact that the number of templates required to achieve a predetermined template bank efficiency grows significantly as we approach the low-mass regime (roughly as $N_templates∝ℳ_min^-8/3$ [ref] ).

In Pursuit of Love: First Templated Search for Compact Objects with Large Tidal Deformabilities in the LIGO-Virgo Data · Search Parameter Space · 0.97

Cites it as background 29 papers

The two-dimensional metric for the inspiral portion of the signal associated with GW from BNS was originally computed at 1.0 PN order [ref] and then extended to 2.0 PN [ref] .

Parameter space metric for 3.5 post-Newtonian gravitational-waves from compact binary inspirals · Introduction · 0.93
The mismatch can be approximated using a second-order Taylor expansion, allowing to define an analytic metric on the waveform manifold [ref] , which in turn is used to define a regular lattice for placing templates in a locally flat space [ref] .

A geometric template bank for the detection of spinning low-mass compact binaries with moderate orbital eccentricity · Matched Filtering · 0.85
Gravitational-wave detection searches for BBHs operate by matched-filtering the detector data against a bank of modeled waveform templates [ref] .

Template Banks for Binary black hole searches with Numerical Relativity waveforms · Conclusions · 0.97

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] [23]

B. J. Owen, Phys. Rev. D 53, 6749 (1996)

View on Pith 1996

[2] [32]

B. J. Owen, L. Lindblom, C. Cutler, B. F. Schutz, A. Vecch io, and N. Andersson, to appear in Phys. Rev. D (gr-qc/980404 4)

View on Pith

[3] [1]

A. Abramovici et al. , Science 256, 325 (1992). 14

View on Pith 1992

[4] [2]

B. Caron et al. , Class. Quantum Grav. 14, 1461 (1997)

View on Pith 1997

[5] [3]

H. L¨ ucket al. , Class. Quantum Grav. 14, 1471 (1997)

View on Pith 1997

[6] [4]

K. Tsubono, in Gravitational Wave Experiments (World Scientiﬁc, Singapore, 1995), p. 112

View on Pith 1995

[7] [5]

R. Narayan, T. Piran, and A. Shemi, Astrophys. J. 379, L17 (1991); E. S. Phinney, Astrophys. J. 380, L17 (1991). In the intervening years astronomers have estimated event rates b oth lower (based on statistics of the few known binary pulsar s) and higher (based on theoretical evolutions of progenitor p opulations), but their estimates have begun to conver...

View on Pith 1991

[8] [6]

S. F. P. Zwart and H. N. Spreeuw, Astron. Astrophys. 312, 670 (1996)

View on Pith 1996

[9] [7]

E. P. J. van den Heuvel and D. R. Lorimer, Mon. Not. R. Astro n. Soc. 283, L67 (1996); I. H. Stairs et al., astro-ph/9712296; V. M. Lipunov, K. A. Postnov and M. E. Prokhorov, Mon. Not. R. A stron. Soc. 288, 245 (1997)

View on Pith Pith review arXiv 1996

[10] [8]

B. Barish et al. , LIGO Advanced Research and Development Proposal, Caltech /MIT, 1996 (unpublished)

View on Pith 1996

[11] [9]

L. Blanchet, T. Damour, B. R. Iyer, C. M. Will, and A. G. Wis eman, Phys. Rev. Lett. 74, 3515 (1995)

View on Pith 1995

[12] [10]

H. Tagoshi, M. Shibata, T. Tanaka, and M. Sasaki, Phys. R ev. D 54, 1439 (1996)

View on Pith 1996

[13] [11]

Some physicists are in the habit of calling this optimal ﬁltering method simply “Wiener ﬁltering.” However, in the l arger signal processing community, “Wiener ﬁltering” refers to a n altogether diﬀerent technique. In order to avoid confusio n we call this method “matched ﬁltering.”

View on Pith

[14] [12]

C. W. Helstrom, Statistical Theory of Signal Detection , 2nd edition (Pergamon Press, London, 1968)

View on Pith 1968

[15] [13]

T. A. Apostolatos, Phys. Rev. D 52, 605 (1995)

View on Pith 1995

[16] [14]

S. Droz and E. Poisson, Phys. Rev. D 56, 4449 (1997)

View on Pith 1997

[17] [15]

T. Damour, B. R. Iyer and B. S. Sathyaprakash, Phys. Rev. D 57, 885 (1998)

View on Pith 1998

[18] [16]

R. Balasubramanian and S. V. Dhurandhar, Phys. Rev. D 50, 6080 (1994)

View on Pith 1994

[19] [17]

A. Kr´ olak, K. D. Kokkotas, and G. Sch¨ afer, Phys. Rev. D 52, 2089 (1995)

View on Pith 1995

[20] [18]

T. A. Apostolatos, Phys. Rev. D 54, 2421 (1996)

View on Pith 1996

[21] [19]

B. S. Sathyaprakash, Phys. Rev. D 50, R7111 (1994)

View on Pith 1994

[22] [20]

B. S. Sathyaprakash and S. V. Dhurandhar, Phys. Rev. D 44, 3819 (1991)

View on Pith 1991

[23] [21]

B. S. Sathyaprakash and S. V. Dhurandhar, J. Comput. Phy s. 109, 215 (1993)

View on Pith 1993

[24] [22]

S. V. Dhurandhar and B. S. Sathyaprakash, Phys. Rev. D 49, 1707 (1994)

View on Pith 1994

[25] [24]

B. F. Schutz, in The Detection of Gravitational Radiation, edited by D. Blair (Cambridge University Press, Cambridge, England, 1991), p. 406

View on Pith 1991

[26] [25]

S. D. Mohanty and S. V. Dhurandhar, Phys. Rev. D 54, 7108 (1996)

View on Pith 1996

[27] [26]

S. D. Mohanty, Phys. Rev. D 57, 630 (1998)

View on Pith 1998

[28] [27]

C. Cutler and ´E. ´E. Flanagan, Phys. Rev. D 49, 2658 (1994)

View on Pith 1994

[29] [28]

R. Balasubramanian, B. S. Sathyaprakash and S. V. Dhura ndhar, Phys. Rev. D 53, 3033 (1996); 54, 1869(E) (1996)

View on Pith 1996

[30] [29]

E. Poisson and C. M. Will, Phys. Rev. D 52, 848 (1995)

View on Pith 1995

[31] [30]

P. C. Peters, Phys. Rev. 136, B1224 (1964)

View on Pith 1964

[32] [31]

B. Allen, grasp software package, available at http://www.ligo.caltech. edu/LIGO web/lsc/lsc.html

View on Pith

[33] [33]

T. Nakamura, M. Sasaki, T. Tanaka, and K. S. Thorne, Astr ophys. J. 487, L139 (1997)

View on Pith 1997

[34] [34]

J. H. Taylor, Rev. Mod. Phys. 66, 711 (1994). 15

View on Pith 1994