Limitations in the bandwidth of the seismic data leads to significant problems for exploration and production objectives in the oil industry. Seismic interpretation become difficult or inconclusive especially for deeper targets and thin beds and inversion for rock properties is more uncertain.

One approach some companies offer is the acquisition of broadband seismic. Several acquisition techniques are available to effectively extend the range of frequencies in seismic data. But is the new broadband acquisition going to improve dramatically the frequency content considering the high cost of acquiring a new data set? Also, what to do with legacy data? Dismiss it? It is possible to design a mathematical operator to insert into the data the desirable frequencies which can be mathematically consistent. But are these frequencies consistent with the geology? Does it honour the rock properties in the reservoir? How reliable can it be?

The ARK CLS Frequency Shaping approach to this problem is significantly different. It generates seismic wavelets consistent with log data. Two convolution operators are derived simultaneously which optimally broadens the data at both ends (low and high) and honours available well data in a global sense. The process is intrinsically stable and broadly consistent with known AI and reflectivity behaviour in the area. Construction of the operators is a simple process and implementation can be readily performed within the processing module. As an explicit wavelet is not required this removes an inherently weak link on which more sophisticated processes rely.

With this technique it is now possible to routinely generate two datasets, one which has its amplitude spectrum extended towards the low end and other towards the high end, and which can be use either independently of simultaneously to achieve the objectives of the project.