M-GS® factorial kriging is a powerful technique for Pre-stack Noise Attenuation. It allows removing various types of noise in pre-stack data, even the most complex ones, such as multiples or linear noise with various intensity, orientation and frequency. Depending on the type of noise, the solution can be applied in preferential dimensions of the pre-stack domain. In the gather domain for example, it can be used to filter out random and coherent linear noise (hidden source for gradient generation error) as illustrated below (data from the Norwegian Sea).