Deconvolution of band-limited signals

The inverse filter to a minimum-phase wavelet with infinite bandwidth is often calculated from its autocorrelation by applying the Wiener-Levinson algorithm. physical limitations of seismic data require the wavelet to have a finite bandwidth which violates the minimum-phase condition. An unstable filter results when the Wiener-Levinson algorithm is applied to band-limited data. Although "processing noise" is often added to stabilize the algorithm, this procedure can obscure problems associated with the data instability from the seismic interpreter. This paper presents an alternate approach to the deconvolution of band-limited data by mapping the frequency response to span the full spectrum based on a set of sampling rate changes according to the lower and upper frequency cutoffs.