Some remarks on "On the windowed Fourier transform and wavelet transform of almost periodic functions," by JR. Partington and B. Unalmi

J.R. Partington and B. Unalmis consider in their 2001 paper [J.R. Partington, B. Unaltmis, Appl. Comput. Harmon. Anal. 10 (1) (2001) 45-60] the windowed Fourier transform and wavelet transform as tools for analyzing almost periodic signals. They establish Parseval-type identities and consider discretized versions of these transforms in order to construct generalized frame decornpositions. We have found a gap in the construction of generalized frames in the windowed Fourier transform case; we comment on this gap and give an alternative proof. As for the wavelet transform case, in [J.R. Partington, B. Unalmis, Appl. Comput. Harmon. Anal. 10 (1) (2001) 45-60] the generalized frame decomposition is done only for the simplest wavelet, the Haar wavelet; we show how to construct generalized frame decompositions for a wide family of wavelets. (C) 2004 Elsevier Inc. All rights reserved.