Deterministical and Statistical Projection Filtering for Signal-Preserving Noise Attenuation

(F-x) prediction filtering (Canales, 1984) is a widely used method for random noise reduction. Recently, Soubaras (1994) showed that it is an inconsistent method and that the correct way to extract a predictable signal from noise is to compute the noise by applying the autodeconvolved prediction error filter (PEF) to the data rasher than the PEF itself.