Conversion of Seismic Data to Log-Like Displays
When relating seismic data to stratigraphy was first discussed we noted a data comparison over a North Sea oil field in conventional and wavelet processed form (p. 20). Also, the wavelet processed data was tied to well data by means of a synthetic seismogram (p. 6). We noted an improvement in resolution in the wavelet-processed data and also the lack of multiples compared to the conventionally-processed data. This was due to the fact that the greater resolution in the data improves the velocity analysis results so that they can better discriminate between true event velocities and multiple velocities.
The wavelet used in the synthetic seismogram was the same wavelet that was in the wavelet-processed data. The synthetic seismogram was spliced into the seismic data. Excellent correlation between the synthetic seismogram and the wavelet-processed data was noted. Contrasting this to the match between the synthetic seismogram and the conventionally-processed data we concluded that using the same wavelet in both the seismic data and the synthetic seismogram is necessary in order to obtain the best possible agreement. A later discussion concerning qualitative stratigraphic correlations re-enforced this point (see p. 83).
The next Figure shows a part of a sonic log whose vertical scale has been converted to time. Note that it consists of a rapidly-varying component (high frequency) and a slowly-varying component (low frequency) generally increasing with time. The slowly-varying component, which we shall call the trend, is made up of frequency components too low to be carried by seismic data (typically