Earlier chapters have assumed that the shot and the geophone are located in the same place. The reality is that there is often as much as a 3-km horizontal separation between them. The 3-km offset is comparable to the depth of many petroleum reservoirs.
Offset is another dimension in the analysis of data. At the time of writing, this dimension is often represented in field operations by about 48 channels. No one seems to believe, however, that 48 channels is enough. Recording systems with as many as 1024 channels are coming into use.
The offset dimension adds three important aspects to reflection seismology. First, it enables us to routinely measure the velocity of seismic waves in rocks. This velocity has been assumed to be known in the previous chapters of this book. Second, it gives us data redundancy: it gives independent measurements of quantities that should be the same. Superposition of the measurements (stacking) offers the potential for signal enhancement by destructive interference of noise. Third (a disadvantage), since the offset is nonzero, procedures for migration take on another element of complexity. By the end of this chapter we will be trying to deal with three confusing subjects at the same time--dip, offset, and lateral velocity variation.
Theoretically it seems that offset should offer us
the possibility of
identifying rocks by observing the reflection coefficient as a function
of angle, both for P waves and for P-to-S converted waves.
The reality seems to be that neither measurement can be made reliably,
if at all.
See elswhere in this book
a fuller discussion of converted waves,
an interesting subject for research,
with a large potential for practical rewards.
See also Ostrander  and Tatham and Stoffa .
The reasons for the difficulty in measurement,
and the resolution of the difficulty,
are, however, not the goal of this book.
This goal is instead to enable us to deal
effectively with that which is routinely observable.