The calculated shapes of blocking dips differ from the experimental ones since models for charged particle motion with in crystals assume a perfect lattice. In this paper we suggest that the difference between the transverse energy distribution π(E⊥) of the particles coming out from a real crystal and that calculated in a statistical equilibrum continuum model, corrected for thermal vibration effects by a diffusion equation, accounts for crystal defects and is approximately independent of the distribution of transverse energy before diffusion and of the energy of blocked particles. The extraction of π(E⊥) from experimental dips requires the inversion of an integral equation, that is analy = tically solved in a particular case of practical interest. This would provide a way of analysing nuclear reaction time measurements that makes use of the full shape of the blocking dips. An illustrative example of the method to a blocking lifetime measurement in27A1 (p,α)24Mg resonance reactions is presented.