Conversions of shear waves (S) to compressional waves (P), often analyzed as Sp receiver functions, are useful for studying upper mantle layering, and have been used to map upper mantle seismic impedance interfaces in various tectonic settings. Recently, common conversion point stacking of Sp receiver functions has revealed variations of lithospheric thickness across short horizontal distances. However, compared to Ps receiver functions and reflection, only limited work has been done toward quantifying the interpretability of Sp receiver functions, especially in settings where large lateral structure variations are present. Using the spectral element method, Profs. Karen Fischer and Ved Lekic modeled wave propagation and S-to-P conversion through simple synthetic models with varying velocity interface topography, and wrote up their results in a recently-published paper in GJI. They systematically explored the effects of wave frequency content, seismometer spacing, and illumination geometry on common-conversion point (CCP) stacked Sp receiver functions in settings where velocity interface depth varies laterally. They observed that the resolving power of Sp receiver functions decreases with decreasing frequency content, and that upward deflections of velocity interfaces are more difficult to observe than are downward deflections, an asymmetry that primarily arises due to corner diffractions. Furthermore, they documented that the relationship between the angle of illumination and the orientation of the topography of the velocity interfaces largely determines the apparent interface slope and strongly affects the amplitude of Sp phases in the CCP stacks. Indeed, under certain illumination geometries, strong velocity contrasts across a dipping lithosphere-asthenosphere boundary may not produce detectable Sp phases at the surface. What's more, diffractions arising from corners of interface topography can produce artifacts in CCP stacks that masquerade as mid-lithospheric impedance jumps or drops, as well as gently-sloped sub-lithospheric impedance drops. Lekic and Fischer argue that estimates based on Fresnel zone considerations might, in some cases, underestimate the true resolution, and that they are likely to be only appropriate for situations in which abrupt lateral variations in structure do not produce waveform complexities. Their results imply that the interpretation of Sp receiver functions and CCP stacks is not straightforward and that care must be exercised when inferring the presence or absence of lithospheric velocity interfaces.