Cadzow De-Noise Filter is based on rank reduction of matrix on frequency slices to attenuate the random noise. This can be done in pre and/or post stack data. Even though the parameters are similar to RNA (Random Noise Attenuation) module, the working principle is slight different as far as the Eigen values are concerned. In this case, these Eigen vectors are applied on multiple dimensions (3 or more) unlike the two dimensions in RNA. The other advantage of Cadzow De-Noise Filter is that it can work on flat or dipping events. It preserves the dips of steeply dipping events while attenuating the random noise.
Seismic image with different Eigen values - 1, 4 ,40.
The example of using Cadzow De-Noise Filter in the workflow and parametrization.
Stack section: before noise attenuation (left), after (middle) and difference (right).
Amplitude frequency spectrum: before noise attenuation (red), after (green).
