An early approach to pre-stack-time migration velocity analysis is illustrated in Figure 3. The approach was popularized by Paul Fowler at the Stanford Exploration Project and later at Western Geophysical in Houston, Texas. In part (a) we see the general idea is to perform many constant velocity pre-stack time migrations and collect the resulting information into a cube with axis for time, CDP, and velocity. Each fixed CDP (y in this case) is then analyzed to select the optimum velocity at that location. Figure 3b compares more traditional stacking velocity and DMO based approaches to the one described by (a). The general conclusion is that the constant velocity approach is somewhat better than the other two.