f Analysis of a Coherency-constrained Inversion for the Separation of Blended Data: Discovering the Leakage Subspace

So-called blended or simultaneous source techniques are interesting because they offer the potential for better acquisition sampling, improved illumination and greatly enhanced computational efficiency. The procedure of retrieving data as if they were acquired in the conventional way is called deblending. Several inversion techniques have been proposed for solving this ill-posed problem with the iterative estimation and subtraction algorithm being one of them. This particular method uses a dedicated iteration that integrates a coherency-pass filter. However, fundamental questions regarding the convergence and limitations of this method have not previously been addressed. The notion of the leakage subspace is hereby introduced to aid in this quest.