The shot gather shown in Figure was first processed with the high-amplitude noise removal process shown in chapter to produce the results in Figure . The value of w, the multiplier of the median, was 5 in this case. After the bad samples were removed, the inversion of system () was used to predict the noise and the missing data. The missing data was then added to the known data to produce a full set of data. This full set of data then had the noise subtracted from it to produce the signal section shown in Figure . The results in Figure appear reasonable. The noise is well attenuated and the signal appears strong. Although the signal is not well predicted into the missing data area near zero offset, it has been extended several traces. Figure shows the difference between the original data shown in Figure and the calculated signal shown in Figure . While some signal appears in the difference section, it is weak compared to the noise. Some of this apparent signal may be due to differences in the coupling of each receiver's geophones to the Earth. Irregularities in the reflection amplitudes will be poorly predicted by the signal annihilation filter and will appear in the noise section.
Figure shows the data from Figure with the noise removal applied. Next, Figure shows the data from Figure where the inversion was used to predict the missing data while separating signal and noise, as in the previous Figures. Notice that most of the coherent noise discussed previously was not restored into the signal in Figure . Figure shows the difference between the original data and the noise. Little signal has leaked into the noise section and the results appear satisfactory.