A shortcoming of the standard surface nuclear magnetic resonance measurement, the free-induction decay, is that the meaning of the signal's time-dependence is uncertain. Ideally, the parameter describing the timedependence of the FID, called T2∗, carries a strong link to the geometry of the pore space. However, in the presence of background magnetic field inhomogeneity a second mechanism contributes to T-2∗potentially obscuring the link to pore geometry. To improve the understanding of which mechanism controls T-2∗, an approach involving direct modelling of relaxation during pulse effects is proposed. Numerical studies are presented to demonstrate that the complex inversion of surface NMR data provides the sensitivity to gain insight into the magnitude of T-2 from only FID measurements. Multiple inversions are performed for plausible T-2, given the observed magnitude of T2∗, and inversions providing satisfactory data fit can help constrain T2. Relaxation during pulse effects are also shown to be a contributing factor to difficulties describing the signal phase for complex inversions of surface NMR data.