Much of what has been discussed so far is what we might call early shot-by-shot migration. Early practitioners of the imaging art were forced to use what they had. They did not have access to modern computers, so sorting data into any other order was impossible because it had not been recorded, impossible because the technology of the day was not capable of doing it, or impossible because it was just too difficult and expensive to consider. There wasn't any way to estimate velocities from recorded data, so shot-by-shot event imaging was the only practical approach.
Without redundancy, velocity information was obtained only by trial and error. If, when tested, a given v ( z ) was shown to be in error, a new v ( z ) was selected and used to produce a new stick image. This process was repeated until the result was considered geologically reasonable. This meant that a different flat Earth, vertically varying velocity was being used for each new geologic setting, even if the new location was close to the previous one. It also meant that a large number of different velocity functions might have to be tried before a suitable one was found.
The basic steps in historical shot-by-shot imaging are summarized in the following list:
From a computational viewpoint, this suggests that more than one set of calculations based on Figure 17 might be necessary to produce an accurate stick image of any given horizon of interest.