IN NMR, the sample is placed in the probe and is excited by the coil using RF fields. But, errors creep in due to the inhomogeneous nature of the B1 field and also the errors in the static magnetic field(which are taken care of using shim coils). In some of the research papers I read, they just describe that the pulses they design are robust to the RF in-homogeneities in the B1 field. Is it that this cannot be addressed in an analytical way? Moreover, what does one mean when one speaks about the inhomogeneity? Is it the variation in the strength of the field or the spatial variation of it or both?