CoChat

$H^{\infty}$-Optimal Fractional Delay Filters

Open access
Nagahara, Masaaki, Yamamoto, Yutaka · 2013 · IEEE Transactions on Signal Processing

The DOI matches its Crossref record; 53 of the 53 references it cites were found in a registry or an index; 3 of the 18 sources its claims cite were checked for support.

Abstract

Fractional delay filters are digital filters to delay discrete-time signals by a fraction of the sampling period. Since the delay is fractional, the intersample behavior of the original analog signal becomes crucial. In contrast to the conventional designs based on the Shannon sampling theorem with the band-limiting hypothesis, the present paper proposes a new approach based on the modern sampled-data H∞optimization that aims at restoring the intersample behavior beyond the Nyquist frequency. By using the lifting transform or continuous-time blocking the design problem is equivalently reduced to a discrete-time H∞optimization, which can be effectively solved by numerical computation softwares. Moreover, a closed-form solution is obtained under an assumption on the original analog signals. Design examples are given to illustrate the advantage of the proposed method.

Overview

CoChat Summary

TL;DR

The paper proposes designing fractional delay filters via sampled-data \(H^\infty\) optimization, aiming to restore intersample behavior beyond the Nyquist frequency rather than relying on Shannon bandlimiting. Using lifting, the sampled-data problem is reduced exactly to a finite-dimensional discrete-time \(H^\infty\) problem, with a closed-form solution under a first-order low-pass signal model.

Key findings

  • The sampled-data \(H^\infty\) error system is norm-equivalent to a finite-dimensional discrete-time system: \(\|E\|_\infty = \|E_d\|_\infty\), so the design can be solved by standard discrete-time \(H^\infty\) software.
  • Under a first-order low-pass weighting \(\hat W(s)\) with cutoff \(\omega_c>0\), the optimal filter \(\hat K(z)\) has a closed-form solution depending on fractional delay \(d\) and integer delay \(m\), enabling variable fractional delay filtering.
  • In the design example with \(T=1\) sec, \(D=5.5\) sec (\(m=5\), \(d=0.5\)), and \(\omega_c=0.1\) rad/sec \(\approx 0.016\) Hz, the \(H^\infty\) design gave an \(L^2\) reconstruction error of \(1.34\times10^{-2}\), versus \(2.07\times10^{-2}\) for the conventional discrete-time \(H^2\) design.
  • The \(H^2\)-optimal filter appeared closer to the ideal low-pass filter, but showed much larger high-frequency errors and more ringing at signal edges than the proposed \(H^\infty\)-optimal filter.

Methods

  • The design problem is formulated as a sampled-data \(H^\infty\) optimization: find a digital filter \(K\) minimizing the worst-case error between the ideal delayed-and-sampled signal and the filtered sampled signal, with continuous-time input modeled by a stable strictly proper \(W(s)\).
  • Continuous-time lifting transforms the mixed continuous/discrete error system into a time-invariant discrete-time system with infinite-dimensional input/output spaces; the dual operator \(BB^*\) then yields a norm-equivalent finite-dimensional discrete-time system.
  • A closed-form solution is derived for first-order low-pass \(\hat W(s)\), and numerical examples compare the proposed filter with conventional FIR filters designed by discrete-time \(H^2\) optimization using an impulse-invariant discretization of \(\hat W(s)\).

Limitations

  • The paper states that the closed-form solution assumes \(\hat W(s)\) is a first-order low-pass filter, which “may appear somewhat restrictive,” though it is argued to cover many typical cases and variations via robustness properties.
  • The paper notes that \(H^\infty\)-optimal filters are generally difficult to solve analytically; the closed-form result is only for the first-order assumption.
  • Remark 2 states that as \(\omega_c\to\infty\), the optimal filter coefficients \(a_0(d),a_1(d)\to0\) and \(\|E\|_\infty\to\infty\), meaning that if analog signals contain high-frequency components far beyond Nyquist, the worst-case input becomes more severe and the \(H^\infty\)-optimal filter approaches zero.
  • Evident from the text: the numerical validation is limited to one design example and compares only against discrete-time \(H^2\) design; no broader empirical study or real-world implementation is reported.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1109/tsp.2013.2265678

Cites

53 of 53 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Fractional delay process: (A) a continuous-time signal v ( t ) (top left) is delayed by D > 0 . (B) the delayed signal v ( t -D ) is sampled at t = nT , n = 0 , 1 , . . . . (C) the signal v ( t ) is sampled at t = nT , n = 0 , 1 , . . . . (D) digital filtering (fractional delay filter, FDF) to produce (or estimate) the sequence { v ( nT -D ) } from the sampled-data { v ( nT ) } .
Fig. 1 Fractional delay process: (A) a continuous-time signal v ( t ) (top left) is delayed by D > 0 . (B) the delayed signal v ( t -D ) is sampled at t = nT , n = 0 , 1 , . . . . (C) the signal v ( t ) is sampled at t = nT , n = 0 , 1 , . . . . (D) digital filtering (fractional delay filter, FDF) to produce (or estimate) the sequence { v ( nT -D ) } from the sampled-data { v ( nT ) } .From the PDF
Fig. 2 Error system E for designing fractional delay filter K . (A)-(D) correspond to those in Fig. 1.
Fig. 2 Error system E for designing fractional delay filter K . (A)-(D) correspond to those in Fig. 1.From the PDF
Fig. 3 Lifted error system E .
Fig. 3 Lifted error system E .From the PDF
Image 4
From the PDF
Image 5
From the PDF
Fig. 6 Bode plot of ˆ W ( s ) (solid) and its impulse-invariant discretization ˆ W d ( z ) (dash). The vertical line shows the Nyquist frequency.
Fig. 6 Bode plot of ˆ W ( s ) (solid) and its impulse-invariant discretization ˆ W d ( z ) (dash). The vertical line shows the Nyquist frequency.From the PDF
Fig. 7 Bode plot of filters: sampled-data H ∞ design (solid), and discretetime H 2 design (dash).
Fig. 7 Bode plot of filters: sampled-data H ∞ design (solid), and discretetime H 2 design (dash).From the PDF
Fig. 8 Frequency response gain of error system E in Fig. 2: sampled-data H ∞ design (solid), and discrete-time H 2 design (dash).
Fig. 8 Frequency response gain of error system E in Fig. 2: sampled-data H ∞ design (solid), and discrete-time H 2 design (dash).From the PDF
Fig. 9 Time response: sampled-data H ∞ design (above) and discrete-time H 2 design (below) with sampling frequency 1000 (Hz).
Fig. 9 Time response: sampled-data H ∞ design (above) and discrete-time H 2 design (below) with sampling frequency 1000 (Hz).From the PDF
Fig. 10 Absolute value of reconstruction error: sampled-data H ∞ design (above) and discrete-time H 2 design (below).
Fig. 10 Absolute value of reconstruction error: sampled-data H ∞ design (above) and discrete-time H 2 design (below).From the PDF

Verification

Claims the paper makes

3 Valid · 2 Not valid · 16 Not checked · 45 Own results and interpretations

23 more claims

Create a free account to see all 66 claims and to verify the 3 references still open.

See all claims — free account
  • It may state two separate claims as one. 13
  • The quote only partly states this claim. 4
  • Probably not a substantial claim. 26

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 53 references

Research works printed but not linked: 3

  1. [49] D. Donoho, A. Maleki, and M. Shaharam, 'Wavelab 850,' 2006.
  2. 0, MATLAB Codes for H-infinity Optimal Fractional Delay Filters
  3. [43] W. J. Rugh, Linear Systems Theory . Prentice Hall, 1996. -

Cited by

Metadata

Authors
Nagahara, M., & Yamamoto, Y.
Affiliations
Kyoto University (JP)
Venue
IEEE Transactions on Signal Processing
Published
2013-09
Volume
61
Issue
18
Pages
4473-4480
ISSN
  • 1053-587X (ISSN-L)
  • 1941-0476
OpenAlex
W2036326049
Cited by
15
References
53
Citations by year
2024: 1 · 2021: 1 · 2019: 1 · 2018: 2 · 2017: 1 · 2016: 3 · 2015: 3 · 2014: 2
FWCI
4.10
Type
journal article
Language
en
Open access
Locations
  • IEEE Transactions on Signal Processing (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Funders
Japan Society for the Promotion of Science (24560543)
Topics
  • Physical Sciences › Computer Science › Signal Processing › Digital Filter Design and Implementation
  • Physical Sciences › Engineering › Control and Systems Engineering › Stability and Control of Uncertain Systems
  • Physical Sciences › Engineering › Control and Systems Engineering › Advanced Control Systems Design
Keywords
fractional delay filter, H-infinity optimization, sampled-data systems, intersample behavior, analytical solution, digital filter design

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.