I would like to follow up on this old request. I don’t think this request has been addressed, as I cannot find any option in the RBMC job type to define the extraction box size.
Currently, I assume that it is the physical extent of input reference volume that determines the extraction box? In my workflow, I initially extract with a box which is fairly tight around the particles, since the early steps of processing does not need to recover the delocalised high-resolution signal far away from the edge of the particle.
Example:
Detector pixel size 0.69 Å
Longest particle diameter 280 Å, average defocus 1.0 um
Initial extraction 512 pixels (353 Å) cropped to 256 pixels (1.38Å/pixel)
Final extraction 600 pixels (414 Å) cropped to 512 pixels (0.81Å/pixel)
Later stages will need more room around the particles as the reconstruction improves. So, I would like to avoid re-extracting the particles with a larger box (600 pixel box in the example) before RBMC, since passing over all the micrographs an extra time is the slowest type of operation and a 2.8 Å reconstruction should in my experience be sufficient for accurate RBMC.
Would it would, as a work-around, to zero-pad and the reference volume in real-space to reach the desired box extent? And perhaps also zero-pad in Fourier-space to interpolate a smaller pixel size?