The current version of IEC 60060-1 defines that overshoot or oscillations on lightning impulses with frequencies above 500 kHz should be disregarded, and that a mean curve should be drawn for the determination of the test voltage, i.e., the effective peak voltage of the impulse. This definition has not been very practical, and it has led to ambiguity in the parameter evaluation of commonly recorded lightning impulses. In this work we have studied two methods for implementation of gradual limitation of high frequency oscillations by filtering the measured lightning impulses before the parameter evaluation. The first method is using direct filtering of the measured impulse. In the second method a fitted mean curve is first subtracted from the measured curve, and only the remaining residual is filtered. The filtered residual is then added back to the mean curve for parameter evaluation. The characteristics of the filters are based on recent knowledge about the behavior of insulating materials. The aim of this work was to study how well the independent realizations of the two chosen methods lead to coherent results when evaluating both the test voltage and the time parameters.