I heard that the actual dose on the grid is ~20% higher than on detector (Falcon 4i). In regard of this, shall we override the dose during Patch motion correction by providing a 20% bigger value than indicated in the Import Movie job? Thank you!
Hi @andro! Welcome to the forum!
Although you’re right that gold or carbon inside the imaged area will affect the measured dose, typically whatever dose is measured it’s what’s used for image processing. So if you measured 30 e-/Å2/ over vacuum, you’d enter 30.
I hope that’s helpful!
Hi rposert!
I mean that ~20% dose gets lost on the energy filter like Selectrics (not on the grid). This is what I heard from the Fisher Scientific representative working in cryoEM field.
Thank you!
Ah, I see @andro, thank you for clarifying!
Most of the loss from an energy filter is due to inelastic scattering, which should be near zero if you’re measuring dose over vacuum. The losses become more significant when there is ice/sample/carbon between the beam and the detector.
In any case, the general consensus in the field is to use the measured dose even with an energy filter inserted. The input dose is mostly used for dose-weighting – if you are concerned that your input dose may be very far off, you can use Reference Based Motion Correction later during processing (once you have a high-quality map) to estimate empirical dose weights 
I hope that clarifies things!