Short note: Shaping / Matching filters

Mahdi H. Al Mutlaq, Gary F. Margrave

In this paper, we briefly present a review of one of the most common filters in geophysics: that is the matching filter. A matching filter is one of the simplest filters to apply, but one needs to consider two important criteria: i) the matching filter length and ii) an optimum crosscorrelation lag. We provide two examples that analyze the effects of the filter length and the correlation between two dissimilar input traces to be matched. These examples show that an optimum filter length needs to be tested by plotting filter lengths vs residuals norm (L-curve method). In some cases a filter length that is less than the input trace length is ideal and reduces the computational time significantly (Figure 1). In the second example, two dissimilar traces are matched and the result is a larger residual error. The value of this example stresses the importance that both traces to be matched need to be correlatable.