CONCLUSIONS

The results are very satisfying. Naturally an unaliased wavefield is preferable to an aliased wavefield, if the data are interpolated with good fidelity. The interpolation residual is non-zero but mostly small, especially at far offsets. The example presented in this abstract is a 2-D survey, the theory is unchanged in 3-D, and fidelity will probably be better, thanks to the added directions of prediction.