view article
Figure 1
Illustrative comparison of (a) the current movie recording format and (b) the newly developed electron-event representation (EER) format. In the movie format, electron counts are accumulated in fractions of predefined time length that are stored as an image stack. The fraction length must be decided before the experiment and determines the shortest time interval at which beam-induced motion correction can be performed during data processing. Shorter fractions allow finer sampling but also increase the movie file size, necessitating a practical compromise with the fraction length. In the EER format, each electron is recorded with its spatial coordinates and the time of detection. The spatial coordinates have subpixel accuracy with fourfold super-sampling, which allows recovery of information beyond the physical Nyquist periodicity of the detector. During processing, the time interval for beam-induced motion calculations can be selected arbitrarily and optimized for best performance.

IUCrJ
Volume 7| Part 5| September 2020| Pages 780-781
ISSN: 2052-2525
https://doi.org/10.1107/S2052252520011616