# Matched filtering

Matched filtering is a signal processing method that detects a known signal pattern in noisy data by correlating the data with a template of the expected signal, producing a statistic whose signal-to-noise ratio (SNR) is the largest attainable with any linear filter. For a known signal in additive Gaussian noise it is the optimal detector: no other filter gives a higher probability of detection at a fixed false-alarm probability.<sup>[1](https://digitalcommons.uri.edu/ele_facpubs/1288)</sup><sup> • </sup><sup>[2](https://link.springer.com/article/10.12942/lrr-2012-4)</sup> The filter's impulse response is the template reversed in time and conjugated, so its output at the signal's arrival time is the inner product of the template with the data.<sup>[3](https://faculty.nps.edu/rcristi/EO3404/C-Filters/text/3-Matched-Filters.pdf)</sup> It underlies radar and sonar pulse detection, digital communications receivers, image [template matching](https://www.edgepedia.org/template-matching), and the matched-filter searches for gravitational waves from coalescing compact binaries in LIGO, Virgo, and KAGRA data.<sup>[4](https://ar5iv.labs.arxiv.org/html/2107.09378)</sup>

| Key fact | Value |
|---|---|
| Output statistic | Correlation of data with the template, sampled and thresholded; equivalent to the likelihood-ratio test in stationary Gaussian noise<sup>[2](https://link.springer.com/article/10.12942/lrr-2012-4)</sup> |
| Impulse response | Time-reversed, conjugated template: \( h(t) = s(T-t) \)<sup>[5](https://web.mit.edu/16.36/2006directory/lectures/Lectures89.pdf)</sup> |
| Maximum output SNR | \( 2E_{s}/N_{0} \) (continuous time, white noise density \( N_{0} \)); \( E/\sigma^{2} \) in discrete time<sup>[5](https://web.mit.edu/16.36/2006directory/lectures/Lectures89.pdf)</sup><sup> • </sup><sup>[6](https://ocw.mit.edu/courses/6-011-introduction-to-communication-control-and-signal-processing-spring-2010/61ec3ac814f18a1f05b4ce0c90e89282_MIT6_011S10_chap14.pdf)</sup> |
| Processing gain | Equal to the template length \( N \); a 50-sample chirp at 0 dB input SNR gives 17 dB output peak SNR<sup>[3](https://faculty.nps.edu/rcristi/EO3404/C-Filters/text/3-Matched-Filters.pdf)</sup> |
| Colored noise | Pre-whiten the data, then match to the whitened template<sup>[6](https://ocw.mit.edu/courses/6-011-introduction-to-communication-control-and-signal-processing-spring-2010/61ec3ac814f18a1f05b4ce0c90e89282_MIT6_011S10_chap14.pdf)</sup> |
| Gravitational-wave template banks | \( 10^{5} \) to \( 10^{6} \) templates per CBC search; SNR thresholds typically 4 to 8<sup>[7](https://indico.in2p3.fr/event/33553/contributions/145150/attachments/87760/132524/GdR%202024%20matched%20filtering%20review.pdf)</sup> |
| Fast implementation | FFT-based filtering costs \( O(N \log N) \) instead of \( O(N^{2}) \)<sup>[8](https://eprints.gla.ac.uk/266724/1/266724.pdf)</sup> |

## How it works

The filter is derived by maximizing the output SNR of a linear filter sampled at one instant. For a signal \( s(t) \) observed over \( [0,T] \) in additive white Gaussian noise of two-sided density \( N_{0}/2 \), the SNR at the sampler is

\[ \mathrm{SNR} = \frac{\left[ \int_{0}^{T} s(\tau)\, h(T-\tau)\, d\tau \right]^{2}}{\left( N_{0}/2 \right) \int_{0}^{T} h^{2}(T-t)\, dt}, \]

and the [Cauchy–Schwarz inequality](https://www.edgepedia.org/cauchy-schwarz-inequality) bounds this quantity by \( 2E_{s}/N_{0} \), with equality if and only if \( h(t) = s(T-t) \).<sup>[5](https://web.mit.edu/16.36/2006directory/lectures/Lectures89.pdf)</sup> In discrete time the same argument gives an impulse response \( h[n] = s[-n] \) over the template support and zero elsewhere, and detection performance is set entirely by the ratio of signal energy \( E \) to noise variance \( \sigma^{2} \).<sup>[6](https://ocw.mit.edu/courses/6-011-introduction-to-communication-control-and-signal-processing-spring-2010/61ec3ac814f18a1f05b4ce0c90e89282_MIT6_011S10_chap14.pdf)</sup> In the frequency domain the maximizing response is \( H(\Omega) = C\, X^{*}(\Omega)/S_{nn}(\Omega) \cdot e^{-j\Omega n_{0}} \), the conjugated signal spectrum divided by the noise power spectral density; for white noise this reduces to \( h[n] = (C/N_{0})\, x[n_{0}-n] \), and the filter output for a matched template is proportional to the signal's autocorrelation function, peaked at the arrival time.<sup>[9](https://socratic-software.github.io/stochastica/Chap_9.html)</sup>

The correlation interpretation follows directly: the matched-filter receiver is equivalent to multiplying the received signal \( r(t) \) by the template \( s(t) \) and integrating over the template duration.<sup>[5](https://web.mit.edu/16.36/2006directory/lectures/Lectures89.pdf)</sup> The same statistic arises as the likelihood-ratio test for a deterministic signal in stationary Gaussian noise, obtained by correlating the data with the template and comparing the result to a threshold; the optimal SNR is \( \rho := \sqrt{\langle h|h\rangle} \), with \( \rho^{2} = 4 \int_{0}^{\infty} |\tilde{h}(f)|^{2}/S_{n}(f)\, df \).<sup>[2](https://link.springer.com/article/10.12942/lrr-2012-4)</sup> The SNR-maximizing property holds over all linear filters and is independent of the noise probability distribution, but the strongest detection guarantee, via the Neyman–Pearson criterion, requires Gaussian noise.<sup>[2](https://link.springer.com/article/10.12942/lrr-2012-4)</sup><sup> • </sup><sup>[4](https://ar5iv.labs.arxiv.org/html/2107.09378)</sup>

## How it is done

A practical detector has four steps. First, choose the template: the expected signal waveform, or a bank of templates covering the range of unknown signal parameters. Second, whiten the noise when it is colored: a whitening filter \( h_{w}[n] \), which exists whenever the noise PSD is strictly positive, is applied first, and the matched filter is then matched to the whitened pulse \( p[n] = s[n] * h_{w}[n] \), giving \( h_{m}[n] = p[-n] \).<sup>[6](https://ocw.mit.edu/courses/6-011-introduction-to-communication-control-and-signal-processing-spring-2010/61ec3ac814f18a1f05b4ce0c90e89282_MIT6_011S10_chap14.pdf)</sup> Third, compute the correlation, either directly or, for long data, in the Fourier domain: the exact statistic requires inverting the data covariance matrix \( C \), which may not fit in RAM for long signals, so the standard implementation approximates \( C \) with a [circulant matrix](https://www.edgepedia.org/circulant-matrix) and applies the Moore–Penrose pseudo-inverse, setting the spectrum ratio to zero where it falls below a tolerance such as \( 10^{-8} \).<sup>[4](https://ar5iv.labs.arxiv.org/html/2107.09378)</sup> Fourth, threshold the output peaks, typically at \( u \) times the standard deviation of the correlation sequence with \( u \) in the range 3 to 5.<sup>[4](https://ar5iv.labs.arxiv.org/html/2107.09378)</sup>

False-alarm control needs care when the signal position is unknown. The specific false-alarm probability is computed from [order statistics](https://www.edgepedia.org/order-statistic) of the largest peak, and in imaging applications the Gaussian-based peak false-detection estimate \( \Phi(u) \) can be about 30 times smaller than the correct peak-PDF-based value \( \Psi(u) \) at threshold \( u = 4 \).<sup>[4](https://ar5iv.labs.arxiv.org/html/2107.09378)</sup><sup> • </sup><sup>[10](https://www.aanda.org/articles/aa/full_html/2016/05/aa27463-15/aa27463-15.html)</sup> In gravitational-wave searches, false-alarm rates are estimated empirically by time-sliding data from different detectors against each other; a false-alarm rate below about 1 per 100 years marks a candidate unlikely to arise from noise.<sup>[7](https://indico.in2p3.fr/event/33553/contributions/145150/attachments/87760/132524/GdR%202024%20matched%20filtering%20review.pdf)</sup>

## Origin

The method emerged from wartime radar reception research. J. H. Van Vleck and [David Middleton](https://www.edgepedia.org/david-middleton-physicist) introduced the filter in "A Theoretical Comparison of the Visual, Aural, and Meter Reception of Pulsed Signals in the Presence of Noise," published in the [Journal of Applied Physics](https://www.edgepedia.org/journal-of-applied-physics) in 1946; P. M. Woodward's 1951 treatment of radar receivers discusses the same filter, apart from a scaling factor \( 2/N_{0} \), crediting Van Vleck and Middleton's paper.<sup>[11](https://doi.org/10.1063/1.1707666)</sup><sup> • </sup><sup>[1](https://digitalcommons.uri.edu/ele_facpubs/1288)</sup><sup> • </sup><sup>[12](https://www.norbertwiener.umd.edu/crowds/documents/Woodward51.pdf)</sup> G. Turin's tutorial "An introduction to matched filters" (1960, IEEE Transactions on Information Theory), whose proof approach most textbooks follow, and Theodore G. Birdsall's frequency-domain proof "On Understanding the Matched Filter in the Frequency Domain" (1976, IEEE Transactions on [Education](https://www.edgepedia.org/education)) became the standard expositions.<sup>[13](https://doi.org/10.1109/tit.1960.1057571)</sup><sup> • </sup><sup>[14](https://doi.org/10.1109/te.1976.4321082)</sup><sup> • </sup><sup>[15](https://journals.sagepub.com/doi/full/10.1177/0020720918815997)</sup>

## Variants

Several named forms adapt the basic filter. The normalized matched filter thresholds the normalized inner product \( |\langle x/\|x\|, y/\|y\| \rangle| \), bounded to \( [0,1] \) and maximized when the two signals have the same shape; for non-time-limited targets this reduces to normalized cross-correlation.<sup>[16](https://eng.libretexts.org/Bookshelves/Electrical_Engineering/Signal_Processing_and_Modeling/Signals_and_Systems_%28Baraniuk_et_al.%29/13%3A_Capstone_Signal_Processing_Topics/13.04%3A_Matched_Filter_Detector)</sup> The generalized or whitened matched filter handles colored noise through the \( G^{*}(f)/|H_{N}(f)|^{2} \) response described above.<sup>[17](https://www.lntwww.lnt.ei.tum.de/Theory_of_Stochastic_Signals/Matched_Filter)</sup> The widely linear matched filter extends the strictly linear filter to improper (noncircular) complex noise by using both the covariance and pseudo-covariance matrices; for proper noise the strictly linear filter's SNR is exactly half of the widely linear one.<sup>[18](https://arxiv.org/pdf/2401.16729v2.pdf)</sup> When signal parameters are unknown, a bank of matched filters runs one filter per candidate template; gravitational-wave template banks place templates on a hexagonal \( A_{2}^{*} \) lattice so the mismatch distribution is essentially flat between 0 and the maximum.<sup>[19](https://arxiv.org/html/1303.2005)</sup>

## Applications

Radar and sonar receivers use matched filters to detect pulse echoes and estimate time of arrival, and communications receivers use them as the optimal detectors in Gaussian noise.<sup>[3](https://faculty.nps.edu/rcristi/EO3404/C-Filters/text/3-Matched-Filters.pdf)</sup><sup> • </sup><sup>[16](https://eng.libretexts.org/Bookshelves/Electrical_Engineering/Signal_Processing_and_Modeling/Signals_and_Systems_%28Baraniuk_et_al.%29/13%3A_Capstone_Signal_Processing_Topics/13.04%3A_Matched_Filter_Detector)</sup> The most prominent modern application is gravitational-wave detection. The FINDCHIRP algorithm, used by LIGO Scientific Collaboration and Virgo searches for coalescing compact binaries, implements the optimal filter with innovations for unknown signal parameters and nonstationary, non-Gaussian detector artifacts.<sup>[20](https://doi.org/10.1103/physrevd.85.122006)</sup> Current LIGO/Virgo/KAGRA compact-binary searches run four template-based matched-filter pipelines, PyCBC, GstLAL, MBTA, and SPIIR; the GstLAL bank for the O4 observing run contains about \( 2 \times 10^{6} \) templates, and GstLAL played a key role in the detection of GW150914 and was the first pipeline to detect GW170817 in low latency.<sup>[7](https://indico.in2p3.fr/event/33553/contributions/145150/attachments/87760/132524/GdR%202024%20matched%20filtering%20review.pdf)</sup><sup> • </sup><sup>[21](https://arxiv.org/html/2410.16416)</sup>

## Limitations and alternatives

The method's central limitation is that it requires the template's functional form: when the shape of the expected signal is not available, there is no general procedure to construct a matched filter.<sup>[4](https://ar5iv.labs.arxiv.org/html/2107.09378)</sup> Mismatch degrades performance quantitatively: any non-matched filter meeting the Nyquist zero-interference criterion loses a factor \( G^{2} \) in SNR equal to the squared normalized cross-correlation between pulse and filter, about 0.967 for a raised-cosine pulse with roll-off 0.5.<sup>[15](https://journals.sagepub.com/doi/full/10.1177/0020720918815997)</sup> In compact-binary searches, template banks are sized so the worst-case fractional SNR loss stays small; a minimal match of 97% ensures less than 10% of signals at the worst-mismatch locations are lost.<sup>[19](https://arxiv.org/html/1303.2005)</sup> Even a 3% template mismatch can produce a large chi-squared value, so FINDCHIRP thresholds a modified, mismatch-weighted chi-squared statistic rather than a fixed one, and modern pipelines use time-frequency, autocorrelation, and bank chi-squared discriminants to reject non-Gaussian artifacts.<sup>[20](https://doi.org/10.1103/physrevd.85.122006)</sup><sup> • </sup><sup>[7](https://indico.in2p3.fr/event/33553/contributions/145150/attachments/87760/132524/GdR%202024%20matched%20filtering%20review.pdf)</sup> For continuous-wave searches the parameter space is so large that directly implemented optimal matched filtering is not computationally feasible, motivating hierarchical searches that are deliberately, and acceptably, sub-optimal.<sup>[22](https://ar5iv.labs.arxiv.org/html/gr-qc/9912029)</sup>

Machine-learning detectors are a recent alternative, and the relationship is close: matched filtering with a collection of templates is formally equivalent to a particular shallow feedforward network, which can be constructed analytically from the templates and then trained on data; deeper variants handle non-Gaussian noise where the optimal decision region is nonconvex.<sup>[23](https://ar5iv.labs.arxiv.org/html/2104.03961)</sup> Hybrid designs feed an SNR map, a stack of matched-filter SNR time series from 256 templates, into a neural network, reaching sensitivity comparable to PyCBC.<sup>[24](https://arxiv.org/pdf/2512.12399)</sup>

Computation scales with the number of templates. Direct time-domain correlation over \( M \) samples costs \( O(M^{2}) \), reduced to \( O(M \log M) \) by the fast [Fourier transform](https://www.edgepedia.org/fourier-transform); gravitational-wave banks can reach about \( 10^{12} \) templates costing about \( 10^{6} \) CPU hours classically, which motivates the quantum Grover search proposed by Sijia Gao and colleagues in [Physical Review Research](https://www.edgepedia.org/physical-review-research) in 2022, with speedup proportional to the square root of the template count.<sup>[8](https://eprints.gla.ac.uk/266724/1/266724.pdf)</sup> GPU implementations cut costs sharply: the chi-squared veto runs about 20 times faster on GPU, and a PyTorch adaptation of the GstLAL LLOID engine reaches a speedup of up to 169 times over a single CPU core on an [NVIDIA A100](https://www.edgepedia.org/nvidia-a100).<sup>[20](https://doi.org/10.1103/physrevd.85.122006)</sup><sup> • </sup><sup>[21](https://arxiv.org/html/2410.16416)</sup> Building on the relative binning method of Barak Zackay and colleagues (2018, arXiv), a 2025 ratio-filter method reconstructs the SNR time series of nearby templates from a reference template's series, agreeing with standard matched filtering to \( O(10^{-4}) \) while cutting computational cost by at least about 25%.<sup>[25](https://doi.org/10.48550/arxiv.1806.08792)</sup><sup> • </sup><sup>[26](https://arxiv.org/html/2511.12894)</sup>

## References

1. [The ubiquitous matched filter: A tutorial and application to radar detection (Kay & Rangaswamy)](https://digitalcommons.uri.edu/ele_facpubs/1288)
2. [Gravitational-Wave Data Analysis. Formalism and Sample Applications: The Gaussian Case (Living Reviews in Relativity)](https://link.springer.com/article/10.12942/lrr-2012-4)
3. [Matched Filters (NPS EO3404 course text, R. Cristi)](https://faculty.nps.edu/rcristi/EO3404/C-Filters/text/3-Matched-Filters.pdf)
4. [Everything you always wanted to know about matched filters (but were afraid to ask) (Vio et al.)](https://ar5iv.labs.arxiv.org/html/2107.09378)
5. [Lectures 8 & 9: Signal Detection: matched filter (MIT 16.36, Eytan Modiano)](https://web.mit.edu/16.36/2006directory/lectures/Lectures89.pdf)
6. [Signals, Systems and Inference, Chapter 14: Signal Detection (Oppenheim & Verghese, MIT OCW)](https://ocw.mit.edu/courses/6-011-introduction-to-communication-control-and-signal-processing-spring-2010/61ec3ac814f18a1f05b4ce0c90e89282_MIT6_011S10_chap14.pdf)
7. [Analysis of ground-based detector data with a focus on matched filtering (GdR 2024 review talk, Tito Dal Canton)](https://indico.in2p3.fr/event/33553/contributions/145150/attachments/87760/132524/GdR%202024%20matched%20filtering%20review.pdf)
8. [Quantum algorithm for gravitational-wave matched filtering (Phys. Rev. Research 4, 023006)](https://eprints.gla.ac.uk/266724/1/266724.pdf)
9. [The Matched Filter, Introduction to Stochastic Signal Processing (Ch. 9)](https://socratic-software.github.io/stochastica/Chap_9.html)
10. [The correct estimate of the probability of false detection of the matched filter in weak-signal detection problems (Astronomy & Astrophysics)](https://www.aanda.org/articles/aa/full_html/2016/05/aa27463-15/aa27463-15.html)
11. [J. H. Van Vleck, David Middleton (1946). A Theoretical Comparison of the Visual, Aural, and Meter Reception of Pulsed Signals in the Presence of Noise. Journal of Applied Physics.](https://doi.org/10.1063/1.1707666)
12. [Theory of Radar Receivers (P.M. Woodward, 1951)](https://www.norbertwiener.umd.edu/crowds/documents/Woodward51.pdf)
13. [G. Turin (1960). An introduction to matched filters. IEEE Transactions on Information Theory.](https://doi.org/10.1109/tit.1960.1057571)
14. [Theodore G. Birdsall (1976). On Understanding the Matched Filter in the Frequency Domain. IEEE Transactions on Education.](https://doi.org/10.1109/te.1976.4321082)
15. [An elucidating proof that a matched filter is optimum (Int. J. Electrical Engineering Education)](https://journals.sagepub.com/doi/full/10.1177/0020720918815997)
16. [13.04: Matched Filter Detector (eng.libretexts.org)](https://eng.libretexts.org/Bookshelves/Electrical_Engineering/Signal_Processing_and_Modeling/Signals_and_Systems_%28Baraniuk_et_al.%29/13%3A_Capstone_Signal_Processing_Topics/13.04%3A_Matched_Filter_Detector)
17. [Matched Filter - Theory of Stochastic Signals (LNTwww, TU Munich)](https://www.lntwww.lnt.ei.tum.de/Theory_of_Stochastic_Signals/Matched_Filter)
18. [Widely Linear Matched Filter: A Lynchpin towards the Interpretability of Complex-valued CNNs (2024)](https://arxiv.org/pdf/2401.16729v2.pdf)
19. [The Balancing Act Of Template Bank Construction: Inspiral Waveform Template Banks For Gravitational-Wave Detectors](https://arxiv.org/html/1303.2005)
20. [Bruce Allen and colleagues (2012). FINDCHIRP: An algorithm for detection of gravitational waves from inspiraling compact binaries. Physical review. D. Particles, fields, gravitation, and cosmology/Physical review. D. Particles and fields.](https://doi.org/10.1103/physrevd.85.122006)
21. [Scalable matched-filtering pipeline for gravitational-wave searches of compact binary mergers (2024)](https://arxiv.org/html/2410.16416)
22. [An Efficient Matched Filtering Algorithm for the Detection of Continuous Gravitational Wave Signals](https://ar5iv.labs.arxiv.org/html/gr-qc/9912029)
23. [Generalized Approach to Matched Filtering using Neural Networks (2021)](https://ar5iv.labs.arxiv.org/html/2104.03961)
24. [Hybrid algorithm combining matched filtering and convolutional neural networks for gravitational wave detection (December 2025)](https://arxiv.org/pdf/2512.12399)
25. [Zackay, Barak, Dai, Liang, Venumadhav, Tejaswi (2018). Relative Binning and Fast Likelihood Evaluation for Gravitational Wave Parameter Estimation. arXiv (Cornell University).](https://doi.org/10.48550/arxiv.1806.08792)
26. [Efficient Reconstruction of Matched-Filter SNR Time Series from Nearby Templates for CBC Searches (November 2025)](https://arxiv.org/html/2511.12894)

---
*Topic: Encyclopedia › Technology and the built world › Communications and everyday technology › Receiver signal processing methods*

*Initially written Sep 29, 2026 · Reviewed: Sep 30, 2026 · Edited: Oct 15, 2026 · Last review: Sep 30, 2026*

*Copyright 2026 EdgeChat AI, a subsidiary of Biostate AI.*

License: Edgepedia Community License 1.0, https://www.edgepedia.org/edgepedia/license