If one wants to compute the wavelet transform of a continuous signal, then one has to assume that its digital representation is derived as in Equation (10.9) and Figure 10.9. However, the digital representation of a signal is usually obtained by filtering it with a band-limiting filter and sampling it at or above the Nyquist frequency. Supposing that the continuous signal, in order to be digitized, is filtered with a filter arbitrarily close to the ideal one, find an expression for the mean-square error made in the computation of the wavelet transform. Discuss the implications of this result in the accuracy of signal processing operations carried out in the wavelet transform domain.

Transcribed Image Text:

x(n) = = 26* (21-n)x(t) dt. (10.9) 00