We suggest a new technique for processing head-wave data with a multistage Wiener filter, which provides conversion and stacking of head waves from a receiver array into a user-specified point. Different conversion algorithms are compared and shown to be a combination of parallel and serial Wiener filters.
The Wiener filter accuracy for a conversion iteration depends on the in-line fold and on the coherence spectrum. The latter describes the signal/noise ratio dependence where head waves are the signal and the noise includes all other components and microseisms. Equations for the accuracy of head-wave traces have been derived from a corresponding equation for a single conversion iteration using the multistage Wiener filter schemes and proceeding from the theory of coupling stochastic systems. The obtained equations are applicable to different field data processing algorithms.
The discussed conversion algorithms differ in noise regularization and are fit for data of any quality.