On Spline Wavelets Dedicated to Professor Charles Chui on the occasion of his 65th birthday Jianzhong Wang Abstract. Spline wavelet is an important aspect of the constructive theory of wavelets. This paper consists of three parts. The ¯rst part surveys the joint work of Charles Chui and me on the construction of spline wavelets based on the duality principle. It also includes a discussion of the computational and algorithmic aspects of spline wavelets. The second part reviews our study of asymptotically time- frequency localization of spline wavelets. The third part introduces the application of Shannon wavelet packet in the sub-band decompo- sition in multiple-channel synchronized transmission of signals. x1. Introduction In this paper, I give a brief survey of my joint work with Charles Chui on spline wavelet. In the theory of wavelet analysis, an important aspect is to construct wavelet bases of a function space, say of the Hilbert space L2(R) ([32], [64], [65]). Recall that a (standard) wavelet basis of L2(R) is the basis generated by the dilations and translates of a \wavelet" function Ã: More precisely, if the set of functions j=2 j fÃj;kgj;k2Z ;Ãj;k = 2 Ã(2 x ¡ k); (1) forms an unconditional basis (also called a Riesz basis) of L2(R); then we call à a wavelet and call the system (1) a wavelet basis. Three of most basic properties of a wavelet are its regularity, its decay speed, and the order of its vanishing moments. Therefore, a wavelets is usually chosen to be compactly supported (or exponentially decay) and to have certain degree of regularity so that they are \local" in both x-domain (time domain) and !-domain (frequency domain). Then each element in the Conference Title 1 Editors pp. 1{6. Copyright Oc 2005 by Nashboro Press, Brentwood, TN. ISBN 0-0-9728482-x-x All rights of reproduction in any form reserved. 2 J.Z. Wang wavelet basis (1) has a ¯nite time-frequency window. Thus, wavelet bases are useful tools for \local" analysis of functions. Many books and papers already explain the importance of wavelet bases in harmonic analysis and in various applications. (See [6], [7], [32], [49], [64], [65].) In the earlier time in the history of wavelet analysis, people tried to construct a wavelet basis by ¯nding the wavelet function à directly. It is Mayer and Mallat who create Multiresolution Analysis (MRA) [64], [65], which provides a powerful framework for the construction of wavelet bases. De¯nition 1. A multiresolution analysis of L2 is a nest of subspaces of L2: ¢ ¢ ¢ ½ V¡1 ½ V0 ½ V1 ½ ¢ ¢ ¢ that satis¯es the following conditions. (1) \j2ZVj = f0g, 2 (2) [j2ZVj = L , (3) f(¢) 2 Vj if and only if f(2¢) 2 Vj+1; and (4) there exists a function Á 2 V0 such that fÁ(x ¡ n)gn2Z is an unconditional basis of V0; i.e. fÁ(x¡n)gn2Z is a basis of V0; and there 2 exist two constants A; B > 0 such that, for all (cn) 2 l ; the following inequality holds: ° ° ° °2 X °X ° X A jc j2 · ° c Á(¢ ¡ n)° · B jc j2: (2) n ° n ° n n2Z n2Z n2Z The function Á described in De¯nition 1 is called an MRA generator. Furthermore, if fÁ(x ¡ n)gn2Z is an orthonormal basis of V0, then Á is called an orthonormal MRA generator. Since Á 2 V0 is also in V1; we can establish a two-scale equation for Á: X 2 Á(x) = 2 hmÁ(2x ¡ m); (hm)m2Z 2 l ; (3) m2Z where h = (hm)m2Z is called the mask of Á: Taking the Fourier transform of (3), we obtain X ^ ¡i!=2 ^ m Á(!) = H(e )Á(!=2);H(z) = hmz ; (4) m2Z which represents the two-scale equation of Á in the frequency domain. Here, H(z) is the symbol of Á: The MRA approach to wavelet basis can be described as follows. Let Wj be a complement of Vj with respect to Vj+1: Wj © Vj = Vj+1; j 2 Z: On Spline Wavelets 3 Then we have 2 L (R) = ©j2ZWj;Wj \ Wk = f0g; j 6= k; g 2 Wj () g(2¢) 2 Wj+1; j 2 Z: Let à 2 W0 be such a function that fÃ(¢ ¡ n)gn2Z forms a Riesz basis 2 of W0: Then fÃj;kgj;k2Z forms a wavelet basis of L (R): Since W0 ½ V1; there is a sequence g 2 l2 such that the function à satis¯es X 2 Ã(t) = 2 gkÁ(2x ¡ k); g 2 l ; (5) k2Z where g = (gk) is the mask of Ã: The Fourier transform of à is X ^ ¡i!=2 ^ m Á(!) = G(e )Á(!=2);G(z) = gmz ; m2Z where G(z) is the symbol of Ã: According to the analysis above, once we have an MRA generator Á (and its mask h) to construct a wavelet basis becomes to ¯nd the mask g of Ã: Let us consider the decomposition of a function f 2 L2 into the wavelet series: X f = dj;kÃj;k (6) j;k2Z or X X X f = cJ;kÁJ;k + dj;kÃj;k: (7) k2Z j¸J k2Z To ¯nd the coe±cients (dj;k) and (cj;k) in the decompositions (6) and ~ ~ (7), we need the dual bases of fÁj;kg and fÃj;kg, say fÁj;kg and fÃj;kg, respectively. The theory of the construction of wavelet bases studies the ~ ~ methods to ¯nd fÃj;kg, fÁj;kg; and fÃj;kg from fÁj;kg: Among various wavelet bases, spline wavelet bases play an important role due to their beautiful structure and powerful ability in computa- tion. Spline wavelet is an important aspect of the constructive theory of wavelets. The paper consists of three parts. The ¯rst part surveys the joint work of Charles Chui and me on the construction of spline wavelets based on the duality principle. It also includes a discussion of the com- putational and algorithmic aspects of spline wavelets. The second part review our study of asymptotically time-frequency localization of spline wavelets. The third part introduces the applications of Shannon wavelet packet in the sub-band decomposition in multiple-channel synchronized transmission of signals. As a survey paper, the proofs of all results will not be included. Instead, we only refer the original papers which show the proofs in details. 4 J.Z. Wang x2. Duality Principle and Construction of Spline Wavelets As we mentioned in the introduction, to construct a wavelet basis based on an MRA generator Á and to compute the coe±cients of the wavelet series, we need to ¯nd the relation among Á; Ã; Á;~ and Ã:~ Their relations can be formulated to Duality principle. Charles and I established the duality principle in [16] , [17], and [18]. Its comprehensive description was included in Charles' book [6]. The book and the papers were published more than ten years ago. Since then, a lot of generalizations of the duality principle have been developed for multivariate wavelet bases (see [44], [45], [46], [47], [48], [53], [54], [56], [57], [58], [66], [67], [68]) , multiwavelet bases (see [14], [36], [37]), and wavelet frames (see [8], [9], [10], [12], [34], [70], [71],[72]). The references listed here are far from complete ones. The readers can refer the references in the papers mentioned above. However, today giving a review of the original idea of the duality principle still makes sense. I now introduce the principle of its original formulation in a concise way. De¯nition 2. Assume a scaling function X Á(t) = 2 h(k)Á(2t ¡ k) (8) k2Z generates an MRA fVjgj2Z: If a scaling function X Á~(t) = 2 h~(k)Á~(2t ¡ k) (9) k2Z satis¯es ~ < Á0;n; Á0;m >= ±nm; (10) then Á is called a dual scaling function of Á~. ~ The dual scaling function Á also generates an MRA fV~jgj2Z; called a dual MRA of fVjgj2Z: It is not hard to see that there is a complement of V~j with respect to V~j+1; say W~ j; which satis¯es 2 L (R) = ©j2ZW~ j; W~ j \ W~ k = f0g; j 6= k; g~ 2 W~ j () g~(2¢) 2 W~ j+1; j 2 Z: ~ ~ and there is a wavelet à 2 W~ 0 such that fÃjkgk2Z forms a Riesz basis of W~ j and ~ hÃi;k; Ãj;li = ±ij±kl; for all i; j; k; l 2 Z: (11) ~ ~ We call à a dual wavelet of à and call fÃjkgj;k2Z and fÃjkgj;k2Z are dual bases of L2: There is a sequenceg ~ 2 l2 such that the function Ã~ satis¯es X Ã~(t) = 2 g~(k)Á~(2x ¡ k): (12) k2Z On Spline Wavelets 5 The relations among Á; Á;~ Ã; and Ã~ are presented in the following duality principle [17], [18]. Theorem 1. Let Á and Á~ be the dual scaling functions with masks h = ~ ~ (hk) and h~ = (hk) respectively, and à and à be the dual wavelets de¯ned by (5) and (12) respectively. then P ~ 2 Pk h(k)h(k ¡ 2l) = ±0l; 2 g(k)~g(k ¡ 2l) = ± ; Pk 0l k; l 2 Z; k h(k)~g(k ¡ 2l) = 0; P ~ k h(k)g(k ¡ 2l) = 0; which is equivalent to H(z)H~ (z) + H(¡z)H~ (¡z) = 1; G(z)G~(z) + G(¡z)G~(¡z) = 1; z 2 ¡; (13) G~(z)H(z) + G~(¡z)H(¡z) = 0; G(z)H~ (z) + G(¡z)H~ (¡z) = 0: P P P k ~ ~ k k where HP(z) = h(k)z ; H(z) = h(k)z ;G(z) = g(k)z ; and G~(z) = g~(k)zk: By duality principle, ¯nding Á;~ Ã; and Ã~ from Á; can be completed by ¯nding H~ (z);G(z); and G~(z) from H(z): Here the symbol H~ (z) must satisfy the condition such that it generates a stable scaling function Á:~ We now see how to apply the duality principle to some special cases.