5D interpolation methods have proven to be capable of overcoming the constraints of actual data acquisition in a wide range of situations. However, in general the interpolation methods work by making some type of assumption about the simplicity or sparsity of the underlying model that describes the seismic data, so there is a legitimate concern about whether any information or resolution is being lost during the interpolation process. To address this concern, we present a simple, general method of measuring 5D leakage. 5D leakage is the noise, and possibly the signal, that 5D interpolation is not able to correctly interpolate due to the fact that the data do not completely conform to the simplicity constraints used by the interpolation algorithm. Using the MWNI algorithm, we show with real data examples that 5D leakage can contain complex (quickly varying) aspects of the data such as noise, diffraction patterns and acquisition footprint. In general, the slowly varying aspects of the signal (especially the reflections) are not part of the measured 5D leakage so most important aspects of the signal appear to be reliably interpolated.