ABSTRACT
A subclass of wavelet transforms [6] has an intimate connection with the theory of digital filter banks [7-10]. Filter banks have been known to the signal processing community for over 3 decades (see [7] and references therein). It is this relation that makes it possible to construct in a systematic way a wide family of wavelets with several desirable properties such as compact support (i.e., finite duration), smoothness, good time-frequency localization, and basis orthonormality (all these terms will be explained later).
