(Integral of scaling function) Let is a scaling function of an MRA. Then
By the definition ( Multiresolution analysis )-1,2 we have We calculate the last term using the proposition ( Basic properties of Fourier transform )-3: We use the formula ( Property of scale and transport 6 ): Using the propositions ( Dominated convergence theorem ) and ( Fubini theorem ), The collection is an orthonormal basis in . We use the proposition ( Parseval equality ). We use the proposition ( Basic properties of Fourier transform )-4,5. Thus
(Integral of wavelet) Let is a scaling function of an MRA and is the associated wavelet, see the proposition ( Scaling equation 2 ). We have
According to the proposition ( Scaling equation 2 ), Thus According to the proposition ( Scaling equation 3 ), Thus where According to the proposition ( Scaling equation ), an according to the proposition ( Integral of scaling function ) thus
(Properties of sequences h and g) We assume the setup ( Discrete wavelet transform ), and is real-valued. Then
(a) ,
(b) ,
(c) ,
(d) ,
(e) ,
(f) .
of (a). Within the proof the proposition ( Integral of wavelet ) we saw
of (b). According the proposition ( Integral of wavelet ), Change of variable, and then . We use the proposition ( Integral of scaling function ).
of (c),(d),(e). We use orthonormality of for every : We use the formula ( Property of scale and transport 7 ). The property (d) follows similarly from orthonormality of for every , see the proposition ( Existence of orthonormal wavelet bases 2 ). The property (e) follows by considering , see the proposition ( Existence of orthonormal wavelet bases 1 )-c.
of (f). We combine the proposition ( Recursive relationships for wavelet transform )-c with ( Recursive relationships for wavelet transform )-a,b: We now express and in terms of using ( Recursive relationships for wavelet transform )-a,b. Thus