Progressive Gelfand-Shilov Spaces and Wavelet Transforms

We discuss progressive Gelfand-Shilov spaces consisting of analytic signals with almost exponential decay in time and frequency variables. It is shown that such signals enjoy an additional localization property. We define wavelet transform and inverse wavelet transform in progressive Gelfand-Shilov spaces and study their continuity properties. It is shown that with a slightly faster decay in domain we may control the decay of the wavelet transform independently in each variable.


Introduction
In this paper, we study a class of Gelfand-Shilov spaces and their closed subspaces of analytic signals which are almost exponentially localized in time and frequency variables.Since the support of Fourier transform of such signals is in 0, ∞ , those subspaces are called progressive Gelfand-Shilov spaces, see Definition 2.1.We use the well-known characterization of Gelfand-Shilov spaces, Theorem 2.6.Note that in contrast to the proofs given in 1-4 Theorem 2.6 can be proved without arguments based on the Fourier transform invariance properties, which is of interest when dealing with progressive Gelfand-Shilov spaces, see Remark 2.7.In particular, we show that progressive Gelfand-Shilov spaces enjoy an additional localization property, Proposition 2.8.
Note that the description of progressive Gelfand-Shilov type spaces given by almost exponential decay in time and frequency variables does not contain explicit regularity conditions.
The study of progressive Gelfand-Shilov spaces is motivated by the fact that they form closed subspaces of the Hardy space H 2 R which consists of functions in L 2 R whose with the results from Section 2, give a continuity result, Theorem 3.5.Section 4 is devoted to the inverse wavelet transform, whose decay properties are studied in Propositions 4.1 and 4.2, and a continuity result and the inversion formula are given in Theorem 4.3.In the last section, we give a reformulation of results from Sections 3 and 4 under the assumption that both the analyzing wavelet and the analyzed function have slightly faster decay.This opens the possibility to control the decay of the wavelet transform in each variable separately.We believe that those results are of independent interest.
In order to simplify the exposition, we observe one-dimensional case only.Extension to higher dimensions is straightforward.Throughout the paper, we use standard notation.For example, L p R , • p , 1 ≤ p ≤ ∞, are Lebesgue spaces, f, g denotes the dual pairing of f and g, and the Fourier transform By D a we denote the dilation We adopt the following notation used in the study of pseudo-differential operators cf.17

Progressive Gelfand-Shilov Spaces
Let h, ν, and μ be positive constants such that ν μ ≥ 1.By S μ ν,h R we denote the Banach space which consists of all f ∈ C ∞ R such that the norm and its topology is given by the inductive limit: Definition 2.1.Let h, ν, and μ, be positive constants such that ν μ ≥ 1.The space and its topology is given by the inductive limit: In the study of ultradistributions, it is of interest to observe projective limits of Banach spaces of test functions, apart from the inductive limits.Results of the present paper can be reformulated for progressive Gelfand-Shilov type spaces defined as projective limits of the spaces S μ, ν,h R .For the projective limits of S μ ν,h R , h > 0, we refer to 21, 22 .Note also that the decay of x α f β x L ∞ can be controlled by some other choice of sequences {M α }, {N β }, instead of the Gevrey sequences M α α! ν and N β β! μ used in 2.1 .We refer to 1, 21-23 for the usual conditions on {M α } and {N β }.
For the sake of the clarity of the exposition, in this paper, we observe the inductive limits and the Gevrey sequences α! ν and β! μ , which is not an essential restriction.
It can be shown that multiplication by x and differentiation are continuous operations in S μ, ν R and that the same is true for translation and dilation by a positive factor, for example 2, Chapter IV .On the other hand, S μ, ν R is not closed under modulations.This implies that Short-time Fourier transform, an important tool in local time-frequency analysis see 24 for the definition and basic properties , is not appropriate for the analysis of S μ, ν R .For that reason, we cannot use the elegant arguments from for example 3, 4 in our analysis.
The following proposition can be used in the proof of Theorem 2.6, and it is of independent interest.Proposition 2.3.Let s > 0 and f ∈ C ∞ R .Then, the following conditions are equivalent.a There exists h > 0 such that for some C > 0.
b There exists k > 0 such that The proof is omitted, since it can be found in 17, Proposition 6.1.5., see also 2, Chapter IV .Alternatively, Proposition 2.3 can be proved by the use of the following auxiliary result, see 25 .

2.10
Next, we recall the characterization of Gelfand-Shilov spaces.
Theorem 2.6.Let ν and μ be positive constants such that ν μ ≥ 1 and let f ∈ C ∞ R .Then, the following conditions are equivalent, c There exist h > 0 and C > 0 (which depend only on f) such that d There exist h > 0 and C > 0 (which depend only on f) such that e There exist h > 0 and C > 0 (which depend only on f) such that This theorem is well known, so we omits the proof.It can be proved as in 1, 3, 4 where arguments based on Fourier transform invariance of S ν μ R are used.Another proof which follows from direct computations is given in 17, Theorem 6.1.6., see also 21, 22 .
and f ω o ω n when ω → 0, for every n ∈ N 0 .By Taylor's formula, it follows that where we have used Theorem 2.6 d and n which, together with Theorem 2.6 c , implies that there exist C > 0 and h > 0 such that

Wavelet Transform
In this section, we define the wavelet transform and study its decay and continuity properties.
The wavelet transform W f, g W g f : and can be extended to any dual pair of spaces provided that duality is meaningful.
In particular, W g f is well defined when f, g ∈ S μ ν R .Theorem 2.6 e suggests to observe the behavior of f and g in the configuration and in the frequency domain separately in the study of W g f when f, g ∈ S μ ν R .We introduce following notation for the subexponential weight function: where h, ν ∈ R.
When h > 0 and ν > 1 the weight m h,ν is radial, submultiplicative and subconvolutive.Recall, a nonnegative function for some constant C > 0.
When ν > 1, the following lemma follows from 12, Lemma 7.1 , see also 27 , where it is proved for a general subconvolutive weight.We give here the proof when the weight is subexponential, since similar arguments will be used later on.Moreover, we allow ν 1 here.Lemma 3.1.Let there be given ν ≥ 1, h > 0 and functions f and g such that for some positive constant C.Then, for some C > 0 and for every h such that h ≤ h > 0.

3.7
By the assumption 3.4 , it follows that where we choose h such that h > h > 0. The proof is complete.
Proposition 3.2.Let there be given ν ≥ 1, h > 0 and let f and g satisfy 3.4 for some C > 0.Then, there exist C > 0 such that Proof.When a 1 the so-called "voice transform" W g f b, 1 is just the convolution between f and g, so by Lemma 3.1 we have for every h such that h > h > 0. In particular, when h h 1 a −1/ν h2 −1/ν , we obtain

3.12
When a ∈ 0, 1 , by the symmetry property of the wavelet transform: cf. 9, 2.0.1 and by the first part of the proof we have: The proof is complete.
Proposition 3.2 can be compared to 9 , where an extra factor 1 a −1 appears, and where the different technique is used for the proof.Note that sub exponential growth of m h,ν leads to the convergent integral, whereas the same type of estimate for a weight of polynomial growth of the form x s gives a divergent integral:

R
x a s x −s dx.

3.16
Next, we observe almost exponential decay in the frequency domain and refer to 9, Theorem 12.0.1.for its polynomial decay counterpart.Proposition 3.3.Let there be given μ, h > 0 and let f and g satisfy , ω > 0, 3.17 for some positive constant C > 0.Then, there exist h > 0 and C > 0 such that Proof.By the Parseval's formula, it follows that The change of variables aω ω gives

3.20
When a > 1, we have where 1/a a .Therefore, it is enough to estimate |W g f b, a | when 0 < a ≤ 1.
We split W g f in three parts:

3.25
Next, we make the change of variables: 1/ω ω in I 2 and I 3 and then rename ω into ω.
It remains to estimate I 3 .We have
We end the section with a continuity result when f, g ∈ S μ, ν R .We refer to 9, Theorem 19.0.1 for the case of progressive functions of polynomial decay, and to 19 for the Gelfand-Shilov spaces of type S. First, we introduce the space S μ ν R × R .Definition 3.4.Let there be given positive constants h, ν and μ.The space for some C > 0.
Then S μ ν R×R h>0 S μ ν,h R×R , and the topology is defined by the inductive limit:

3.29
Theorem 3.5.Let there be given ν ≥ 1 and μ > 0 and let f, g ∈ S μ, ν R .Then, the wavelet transform for s max {ν, μ} and for some positive constants h and C. By Propositions 3.2 and 3.3, it follows that there exist positive constants C and h such that which proves the continuity.
Note that f and g in Theorem 3.5 belong to the same progressive Gelfand-Shilov space, while in 19 different decay conditions are imposed on f and g.Since the analyzing wavelet g in 19 may not satisfy the admissibility condition 4.24 , the results given there are of global nature, and the inversion formula cannot be derived.

Inverse Wavelet Transform
In this section, we study the inverse wavelet transform, or the wavelet synthesis operator M g of function τ ∈ S μ ν R × R with respect to the wavelet g ∈ L 1 R given by Our first result is analogous to Proposition 3.2 see also 9, Theorem 18.2.1 .
for some positive constants h and C, then for some positive constants h and C.
Proof.We have where in I 1 the integration with respect to a is taken over 0, 1 and in I 2 over 1, ∞ .By 3.6 , it follows that for any h ∈ 0, h we have for some C > 0. Next, we estimate I 2 .We have

4.9
By Lemma 3.1, x da, x ∈ R, 4.10 for every h ∈ 0, h .After the change of variables a 1/a and renaming a to a, by the mean value theorem, we obtain for some θ ∈ 0, 1 .Therefore, there are C > 0 and h > 0 such that x , x ∈ R, and 4.3 is proved.
The next result can be compared to 9, Theorem 18.1.1 .Here, the assumption on the decay of g leads to a more precise estimate for M g τ then the one given in 9 .for some positive constants h and C, then for some positive constants h and C.
Proof.The Fourier transform of M g τ is given by Therefore,

4.15
where and we have made the change of variables b/ 1 a b .We use similar arguments as in proof of Proposition 3.3 to estimate I 1 ω when 0 < ω ≤ 1.This gives where we have used the inequality ω 1/ω /2 < 1/ω, ω ∈ 0, 1 .
When ω > 1 by the mean value theorem, we obtain where θ ∈ 0, 1 and h θ/2 1/μ h.It remains to estimate I 2 ω , ω > 0. When ω ∈ 0, 1 , by change of variables 1/a a and by the mean value theorem, it follows that where θ ∈ 0, 1 and h θ 1/μ h.In a similar way, as in the proof of Proposition 3.3, it follows that and the proof is complete.

Journal of Function Spaces and Applications
Finally, we have the following result, which can be compared to the inversion formulas from 14, 15 which hold for general coorbit spaces.As mentioned in the introduction, our approach is different due to the fact that the weight m h,ν is not submultiplicative on the "ax b"-group.
, where s max{ν, μ}.Moreover, if g / 0, then for any f ∈ S s, ν R the following inversion formula holds: From the the proof of Proposition 2.8, it follows that g satisfies the admissibility condition: and by Theorem 3.5, we have that W g f ∈ S s ν R × R , from the first part of the proof it follows that M g W g f ∈ S s, ν R .The proof is now the same as the proof of 9, Theorem 14.0.2 if we take g to be its own reconstruction wavelet and observe S s, ν R instead of the space of polynomially strip localized functions.

Concluding Remarks
We first note that under the assumptions of Proposition 3.2 we have that for any given M > 0 there exist h > 0 and C > 0 such that Note that those estimates imply that we are not able to separate variables b and a in a straightforward manner and lead to 3.28 in Definition 3.4.It is probably more natural to have a possibility to analyze decay with respect to variables b and a separately.Therefore, we have the following definition.
, 20 : • 1 | • | 2 1/2 .For reader's convenience, below we give the notation for spaces of test functions introduced in this paper and refer to corresponding definitions.Progressive Gelfand-Shilov spaces S μ, ν R are introduced in Definition 2.1.The range of the wavelet transform over such spaces is contained in S μ ν R × R , see Definition 3.4, and in Z μ ν R × R , see Definition 5.1, the latter being introduced for the analysis of decay of wavelet transform with respect to variables a and b separately, for example, Section 5.

Remark 2 . 7 .Proposition 2 . 8 .
By Theorem 2.6 immediately follows the characterization of progressive Gelfand-Shilov spaces S μ, ν R where the supremum with respect to ω in Theorem 2.6 c , d , and e is taken over ω > 0 instead of ω ∈ R.Moreover, elements in S μ, ν R satisfy an additional localization property, which is sometimes called "strip localization."We refer to 9 for the definition of polynomial strip localized progressive functions.Let us show that elements from S μ, ν R are almost exponentially strip localized.Let ν and μ be positive constants such that ν μ ≥ 1 and let

Proposition 4 . 2 .
Let ν and μ be positive constants.If τ ∈ S μ ν R × R and if g satisfies g ω ≤ C m −1 h

Definition 5 . 1 .Theorem 5 . 2 .
Let there be given positive constants h, ν, and μ.The spaceZ μ ν,h R×R consists of all τ ∈ C ∞ R × R such that |τ b, a | ≤ Cm −1 h,ν b m R × R, and the topology is given by the inductive limit: reformulate results from Sections 3 and 4 in terms ofZ μ ν R × R since we believe that this might be of independent interest.Let ν > 1 and μ > 0, let f, g ∈ S μ, ν−ε R , with ε > 0 chosen such that ν − ε ≥ 1.Then, the wavelet transform W : f, g → W g f is a continuous map from S μ, ν−ε R × S μ, ν−ε R into Z s ν R × R, where s max {ν, μ}.Proof.It follows from the proof of Theorem 3.5 and the fact that for every positive N, M, and ε there is a constant C > 0 such thate −N|•| 1/ ν−ε ≤ Ce −M|•| 1/ν , 5.3 wherefrom W g f b, a ≤ C e h|b/ 1 a | 1/ ν−ε ≤ C e k|b| 1/ν , b, a ∈ R × R , 5.4for arbitrary k > 0. The theorem is proved.The inverse wavelet transform M g of τ ∈ Z μ ν R × R with respect to the wavelet g ∈ L 1 R given by 1 and 4.2 still hold true with S μ ν R × R replaced by Z μ ν R × R , while Theorem 4.3 takes the following form.