Results

The leftmost panel in Figure 1 shows the number of input points contributing to each output point. The shape of the coverage is similar for all three analysis methods, so this shouldn't be an explanation for different results. The amount of coverage is lower for velocity-push because of the fixed computation cost mentioned earlier. However, this lower amount doesn't have a major effect.

Raw semblance and mean-weighted semblance are also displayed. Both measure hyperbolic coherency, while the latter takes event strengths into account. The interesting result is the lack of lateral streaks in the velocity-push results. Lateral streaks are caused by the near offset truncation. For involved reasons inner offsets don't contribute as much data to the velocity analysis as the other methods and truncation artifacts are reduced.