Figures 6 and 13 are the results by GSI
using proprietary software in 1981 and the results obtained using SU.
Some of the differences in the processing sequences are:
The 1981 processing used GSI's designature process. This source
signature removal technique applies the same operator to all traces in a
shotrecord. The filter is designed differently from conventional
deconvolution and the two types of deconvolution will not produce data
with the same phase.
There are many differences in the application of AGC. For example
GSI applied AGC before velocity filtering and inverse AGC after velocity
filter.
GSI used diversity stack to control the high amplitude noise bursts.
Considering all these differences, I think the results are surprisingly
similar, especially below 400 ms. The shallow results are better on the
1981 processing.
Open Data/Open Source: Seismic Unix scripts to process a 2D land line
![[pdf]](icons/pdf.png){kind=icon}