×

Functions of variable bandwidth via time-frequency analysis tools. (English) Zbl 1223.42022

Functions of variable bandwidth are defined as elements of certain modulation spaces. The latter consist of functions or distributions \(f\) whose short-time Fourier transforms \(V_g f(x,\omega)= \int f(t) \, e^{-2\pi i \omega t}\overline{g (t-x)}\, dt = \langle f, \pi (x,\omega) g\rangle\), using a window \(g\), belong to the weighted mixed-norm space \(M_m^{p,q}(\mathbb{R})\) normed by
\[ \|F\|_{M_m^{p,q}} = \biggl(\biggl( \int_{-\infty}^\infty |F(x,\omega)|^p m(x,\omega)^p \, dx\biggr)^{q/p} \, d\omega\biggr)^{1/q} . \]
Here, \(m\) is assumed to satisfy a certain moderate growth condition that allows one to conclude certain containment properties of the spaces, make conclusions about the equivalence of modulation spaces having corresponding equivalent weights, and allow some flexibility in the choice of the window function \(g\). The exponents \(p,q\) are taken in the range \([1,\infty)\) or, properly modified, \(p\) or \(q=\infty\), to make \(M_m^{p,q}(\mathbb{R})\) a Banach space. This work deals with particular weights of the form \(m_{b,s}(z) = (1+d_b(z))^s\), where \(s>0\) and, for \(z=(x,\omega)\), \(d_b(z)= 0\) if \(|\omega|\leq b(x)\) and \(d_b(z)= |\omega-b(x)|\) if \(|\omega|> b(x)\). Here, \(b(x)\geq 0\) is regarded as a variable bandwidth indicator. The space \(\text{VB}^{p,q}_{m_{b,s}}\) is defined as those \(f\) whose short-time Fourier transforms \(V_g f(x,\omega)\) belong to \(M_m^{p,q}(\mathbb{R})\) with \(m=m_{b,s}\). It is shown that \(\text{VB}^{p,q}_{m_{b,s}}\) increases as the exponents \(p\) or \(q\) increase, decreases as the growth index \(s\) increases, and increases as the bandwidth parameter \(b(x)\) increases pointwise. For the particular case \(p=q=2\), it is shown how to build elements of \(\text{VB}^{2,2}_{m_{b,s}}\) by patching together a sequence of bandlimited functions in \(L^2\) having different bandwidths via a partition of unity. To provide a sense in which the \(\text{VB}\) spaces are a natural way in which to quantify variable bandwidth, approximation error estimations are provided. Given a dense enough time-frequency lattice \(\Lambda\) and appropriate \(g\), the operator \(Sf=\sum_{\lambda\in\Lambda} \langle f,\pi(\lambda) g\rangle \pi (\lambda) g\) is continuously invertible in \(L^2\) and \(\gamma=S^{-1} g\), called the canonical dual of \(g\), satisfies \(f= \sum_{\lambda\in\Lambda} \langle f,\pi(\lambda) \gamma \rangle \pi (\lambda) g\). For the case of \(\text{VB}^{2,2}_{m_{b,s}}\)-spaces, one has that, if \((g,\gamma)\) define a \(\Lambda\)-dual pair of windows in the weighted modulation space \(M_\nu^{1,1}(\mathbb{R})\) and \(0<a<s\), then for any \(\varepsilon>0\) there exists \(r>0\) such that for all \(f\in \text{VB}^{2,2}_{m_{b,s}}\), one has the approximation estimate
\[ \biggl\|f-\sum_{\Lambda\cap ST_{b+r}} \langle f, \, \pi(\lambda)\gamma\rangle \pi(\lambda) g\biggr\|_{\text{VB}^{2,2}_{m_{b,s}}} \leq \varepsilon \|f\|_{\text{VB}^{2,2}_{m_{b,s}}}. \]
Here, \(ST_{b+r}=\{(x,\omega):\, |\omega|\leq b(x)+r \}\) and \(\Lambda\) is a discrete lattice. Corresponding approximation results are also proved for more general \(\text{VB}^{p,q}_{m_{b,s}}\)-spaces.

MSC:

42C15 General harmonic expansions, frames
42C40 Nontrigonometric harmonic analysis involving wavelets and other special systems
Full Text: DOI

References:

[1] R. Aceska, Functions of variable bandwidth: a time-frequency approach, PhD thesis, University of Vienna, 2009.; R. Aceska, Functions of variable bandwidth: a time-frequency approach, PhD thesis, University of Vienna, 2009.
[2] R.G. Baraniuk, Warped perspectives in time-frequency analysis, in: Proceedings of the IEEE-SP International Symposium on Time-Frequency and Time-Scale Analysis, October 1994, pp. 528-531.; R.G. Baraniuk, Warped perspectives in time-frequency analysis, in: Proceedings of the IEEE-SP International Symposium on Time-Frequency and Time-Scale Analysis, October 1994, pp. 528-531.
[3] Boudreaux Bartels, G. F.; Parks, T. W., Time-varying filtering and signal estimation using Wigner distribution functions, ISP, 34, 442-451 (1986)
[4] Oppenheim, A.; Braccini, C., Unequal bandwidth spectral analysis using digital frequency warping, IEEE Trans. Acoust. Speech Signal Process., 22, 4, 236-244 (1974)
[5] Casazza, P. G., The art of frame theory, Taiwanese J. Math., 4, 2, 129-201 (2000) · Zbl 0966.42022
[6] Christensen, O., An Introduction to Frames and Riesz Bases, Appl. Numer. Harmon. Anal. (2003), Birkhäuser: Birkhäuser Boston · Zbl 1017.42022
[7] Dabrowska, D., Variable bandwidth conditional Kaplan-Meier estimate, Scand. J. Statist., 19, 4, 351-361 (1992) · Zbl 0768.62024
[8] Doksum, K.; Peterson, D.; Samarov, A., On variable bandwidth selection in local polynomial regression, J. R. Stat. Soc. Ser. B Stat. Methodol., 62, 3, 431-448 (2000) · Zbl 0963.62033
[9] Dubois, E., The sampling and reconstruction of time-varying imagery with application in video systems, Proc. IEEE, 73, 4, 502-522 (1985)
[10] Fan, J.; Gijbels, I.; Hu, Tien-Chung; Huang, Li-Shan, A study of variable bandwidth selection for local polynomial regression, Statist. Sinica, 6, 1, 113-127 (1996) · Zbl 0840.62041
[11] Fan, J.; Jiang, J., Variable bandwidth and one-step local \(m\)-estimator, Sci. China Ser. A, 43, 1, 65-81 (2000) · Zbl 0969.62028
[12] Feichtinger, H. G., Gewichtsfunktionen auf lokalkompakten Gruppen, Sitzungsber. Österr. Akad. Wiss., 188, 451-471 (1979) · Zbl 0447.43004
[13] Feichtinger, H. G., Modulation spaces of locally compact Abelian groups, (Radha, R.; Krishna, M.; Thangavelu, S., Proc. Internat. Conf. on Wavelets and Applications. Proc. Internat. Conf. on Wavelets and Applications, Chennai, January 2002 (2003), New Delhi Allied Publishers), 1-56
[14] Feichtinger, H. G., Minimal Banach spaces and atomic representations, Publ. Math. Debrecen, 34, 231-240 (1987) · Zbl 0562.43003
[15] Feichtinger, H. G.; Gröchenig, K., A unified approach to atomic decompositions via integrable group representations, Lecture Notes in Math., 1302, 52-73 (1988) · Zbl 0658.22007
[16] Feichtinger, H. G., Atomic characterizations of modulation spaces through Gabor-type representations, Proc. Conf. Constructive Function Theory. Proc. Conf. Constructive Function Theory, Rocky Mountain J. Math., 19, 113-126 (1989) · Zbl 0780.46023
[17] Feichtinger, H. G.; Gröchenig, K., Banach spaces related to integrable group representations and their atomic decompositions, I, J. Funct. Anal., 86, 307-340 (1989) · Zbl 0691.46011
[18] Feichtinger, H. G.; Gröchenig, K., Gabor wavelets and the Heisenberg group: Gabor expansions and short time Fourier transform from the group theoretical point of view, (Chui, C. K., Wavelets: A Tutorial in Theory and Applications. Wavelets: A Tutorial in Theory and Applications, Wavelet Anal. Appl., vol. 2 (1992), Academic Press: Academic Press Boston), 359-397 · Zbl 0849.43003
[19] Feichtinger, H. G.; Zimmermann, G., A Banach space of test functions for Gabor analysis, (Feichtinger, H. G.; Strohmer, T., Gabor Analysis and Algorithms: Theory and Applications. Gabor Analysis and Algorithms: Theory and Applications, Appl. Numer. Harmon. Anal. (1998), Birkhäuser Boston: Birkhäuser Boston Boston, MA), 123-170 · Zbl 0890.42008
[20] Feichtinger, H. G.; Werther, T., Robustness of regular sampling in Sobolev algebras, (Benedetto, J. J., Sampling, Wavelets and Tomography (2004), Birkhäuser), 83-113 · Zbl 1070.46022
[21] H. Feichtinger, M. Neuhauser, M. Piotrowski, The homogeneous approximation property and localized Gabor frames, preprint, 2010.; H. Feichtinger, M. Neuhauser, M. Piotrowski, The homogeneous approximation property and localized Gabor frames, preprint, 2010.
[22] M. Greitans, R. Shavelis, Speech sampling by level-crossing and its reconstruction using spline-based filtering, in: 6th EURASIP Conference Focused on Speech and Image Processing, June 2007, pp. 292-295.; M. Greitans, R. Shavelis, Speech sampling by level-crossing and its reconstruction using spline-based filtering, in: 6th EURASIP Conference Focused on Speech and Image Processing, June 2007, pp. 292-295.
[23] Gröchenig, K., Foundations of Time-Frequency Analysis, Appl. Numer. Harmon. Anal. (2001), Birkhäuser Boston: Birkhäuser Boston Boston, MA · Zbl 0966.42020
[24] Gröchenig, K.; Heil, C., Modulation spaces as symbol classes for pseudodifferential operators, (Krishna, R. R.M., Proceedings of International Conference on Wavelets and Applications. Proceedings of International Conference on Wavelets and Applications, 2002 (2003), Allied Publishers: Allied Publishers Chennai, India), 151-170
[25] K. Gröchenig, J. Toft, Isomorphism properties of Toeplitz operators in time-frequency analysis, J. Anal. Math., in press, arXiv:0905.4954v2 [math.FA]; K. Gröchenig, J. Toft, Isomorphism properties of Toeplitz operators in time-frequency analysis, J. Anal. Math., in press, arXiv:0905.4954v2 [math.FA]
[26] Horiuchi, K., Sampling principle for continuous signals with time-varying bands, Inform. and Comput., 13, 1, 53-61 (1968) · Zbl 0162.51503
[27] Slepian, D., On bandwidth, Proc. IEEE, 64, 3, 292-300 (1976)
[28] Wenyang, Z.; Sik Yum, L., Variable bandwidth selection in varying-coefficient models, J. Multivariate Anal., 74, 1, 116-134 (2000) · Zbl 0969.62032
[29] Toft, J., Continuity and Schatten-von Neumann properties for pseudo-differential operators on modulation spaces, (Toft, J.; Wong, M. W.; Zhu, H., Modern Trends in Pseudo-Differential Operators. Modern Trends in Pseudo-Differential Operators, Oper. Theory Adv. Appl., vol. 172 (2007)), 173-206 · Zbl 1133.35110
[30] Weisz, F., Inversion of the short-time Fourier transform using Riemannian sums, J. Fourier Anal. Appl., 13, 3, 357-368 (2007) · Zbl 1138.42015
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.