Adaptive deconvolution by lattice filters: experience on synthetic and real data.

Compares the performances of some types of deconvolution on synthetic and real data, with particular attention to adaptive lattice filters. This technique in fact, besides achieving the desired compression of the seismic signal, anticipates the possibility of a 'continuous' estimation of the basic wavelet by monitoring the lattice coefficients in time. Moreover, some problems connected with the filter adaptivity are discussed, i.e., the sensibility to the adaptation and the introduction of a 'memory', which allows a reduction of the adaptation time and an increase of the statistical robustness of the algorithm. -from Authors