The data we chiefly used was a single 3-component reversed profile. That is, there was a single 3-component receiver fixed in one borehole, and a source which was moved and shot every 50 feet over a span of the other borehole.
At first the data was processed without benefit of precise location information, and there were some traces which had gains that were substantially different to others. In the end, we were able to disentangle the headers and come up with a reasonable idea of the borehole geometries.