The presented interpolation method is based on the common-reflection-surface (CRS) theory. For each sample that has to be interpolated five parameters are required and estimated from the data. Therefore, the proposed method is computational more expensive than existing methods. We present the potential as well as limitations of the method. The result of a CRS interpolation strongly depends on the accuracy of the detected parameters. We analyze the behavior of the estimated parameters for different situations and investigate the effect of an adapted smoothing of the parameters. For the analysis we use the Sigsbee data set which offers numerous complex reflection and diffraction patterns. For a comparison we show the result of a f-x interpolation.