Continuing the discussion from Choose resolution to start applying user-defined mask?:
It would still be useful to have this - particularly as it might allow symmetry relaxation to work using a mask without any additional alterations…
One option would be to change the parameter to “resolution to start masking” - so it applies to the dynamic mask or static mask, as relevant for the specific job?
Oli
Hi @olibclarke,
Thanks for the request. This suggestion makes sense to us; if we change the param to “res to start masking” then the behaviour in the dynamic case wouldn’t change, but the param would still work when a static mask is supplied.
However, below that resolution, no mask will be used (in either the dynamic or static case). Would that be optimal? In any case, I’ve triaged this, thank you!
Best,
Michael
1 Like
However, below that resolution, no mask will be used (in either the dynamic or static case). Would that be optimal? In any case, I’ve triaged this, thank you!
That’s fine - we often use NU-refine with masking disabled completely with good results, sometimes even better than with dynamic or static masking. So having it run without mask for first several iterations is fine I think (with NU-regularization taking care of smoothly downweighting solvent).
My hope is that by disabling masking until some reasonably high resolution, this will allow us to then apply a local, soft, static mask (particularly useful in the context of symmetry relaxed NU-refine), without incurring the mask edge artefacts that appear if such a mask is used from the beginning of refinement.
Thanks Michael!
1 Like