We performed 7 iterations of 3-D geophysical RIP. The result of the stack can be seen in the lower panels of Figure and Figure . The two figures display different inline, crossline, and depth slices, but the types of improvements are the same. In both figures, ovals on the stacked migration result and the stacked geophysical RIP result indicate particular areas where RIP has improved the image. In Fig. , oval ``A'' indicates reflectors that can be followed under the salt nose after imaging with RIP. Ovals ``B'', ``C'', and ``D'' show areas on the inline section where the reflectors can be traced almost entirely through the shadow zones after RIP. In the crossline section, many more reflectors are seen after RIP, particularly in oval ``E''. In Fig. , ovals ``A'',``B'', and ``C'' show areas on the inline section where the reflectors can be traced almost entirely through the shadow zones after RIP and oval ``D'' shows reflectors and a possible fault in the crossline section.
It is not surprising that the comparison of the migration stack and the RIP stack show less impressive improvements than seen in the 2-D example. Performing only 7 iterations of geophysical RIP, which is regularizing only the inline ph axis, will not change the image enough to show very significant effects in the stacked volumes. Although there were more substantial improvements in the stacks after 6 iterations of RIP on a 2-D line taken from this dataset (Fig. ), inverting a 3-D problem means that it can take many more iterations to get similar improvements. However, 3-D RIP does result in clear improvements over the migration results.