Interpolation Error Filters

Claerbout (1989) described the interpolation error filter. This is a filter with a one at zero lag and coefficients either side of zero lag designed such that the power of the output data is minimized.

I.e. we design a filter

$$f={ f_{-n}, f_{-(n-1)} \ldots, f_{-1}, 1, f_1, f_2, \ldots, f_n }$$

to minimize | f * d | 2.