The quality of precession electron diffraction data is higher than necessary for structure solution of unknown crystalline phases.

Four crystal structures are solved from precession electron diffraction data using either the highest-quality experimental data available or voluntarily deteriorated data. The data quality was lowered by changing the intensity of each individual reflection by as much as a factor of 2, even for strong reflections, while taking care to keep strong reflections strong and weak reflections weak. The crystals have been chosen for their different characteristics, ranging from small to large unit cells, low to high symmetry, and containing heavy and light atoms. In each case the structure solution succeeded with the deteriorated data provided that the completeness of the data was high enough. The accuracy of the atom positions obtained for the cations was comparable to that for the best experimental data. Light-atom positions were sometimes less accurate but still satisfactory for a possible subsequent refinement.