Receiver functions generated from distant earthquakes have long been used to determine deep crustal and mantle structure beneath seismic stations. Teleseismic research over sedimentary basins often has to contend with cluttering of the seismic response by basin reverberations. A sedimentary basin typically exhibits a sharp velocity contrast at their base, which acts as a strong reflector for downgoing seismic waves, leading to reverberations within the basin between the basement interface and freesurface. These reverberations obscure important P to S converted arrivals and Moho reverberations in receiver functions. This research tests whether it is possible to remove the effects caused by a sedimentary basin (layer-strip) using imperfect velocity models. A filter is generated using the propagator method, as the spectral ratio of two model responses: one containing the basin and one without. The transfer function, generated as the ratio between the two responses, acts as an acausal filter. Tests on synthetic data were used to evaluate the level of model accuracy required for effective filtering. The method was then tested on real data from Canadian National Seismograph Network station EDM (Edmonton, Alberta, in the Western Canada Sedimentary basin) as well as a line of EarthScope Transportable Array stations over the Williston Basin. Results indicate that the effectiveness of the filter is correlated with the quality of the model, and that the method significantly improves the recovery of Moho reverberations (and thus the quality of an H-k stack) on real teleseismic data collected over sedimentary basins. As a quality basin model is required to achieve trustworthy filtering, the layer-strip methodology is particularly useful in areas where a velocity model is produced with two or more lines of evidence.