Importance of correct direct beam position before first prediction process for the refinement process
The (initial) direct beam position is for existing crystallographic indexing & integration programs a crucial value. A value some pixels off can lead to problems in indexing and refinement (crystal orientation, etc.). The initial direct beam position in NSXTool for the first prediction process is quite a bit off the correct one (see issue #330 (closed). The impact to the following refinement process is quite remarkable. Which I want to show in the following "slide show":
Direct beam position and results of the first "predict":
First refinement result 35 batches, real space:
Second refinement with refined values from first refinement run:
Now we switch back to the "predict" section. After the first (two) refinement runs the direct beam position seems to be updates. It is now correct in respect to the air scattering hallow:
After a second "predict" the results are much better than after the first "predict" cycle (compare first predict):
A subsequent second "refine" run gives great results:
Here for a better view only the refined positions are shown:
Therefore it is important that we get the correct direct beam position quite from the start.