Is it possible to exclude atoms in real-space refinement? Older posts seem to indicate this is, or at least was, only possible in reciprocal-space refinement. Has that changed? Using reference restraints as an alternative is less than optimal. Any help would be greatly appreciated.