mrconvert: fslgrad and RealignTransform: 0

[bjeurissen]

I noticed that running mrconvert data.nii.gz -fslgrad bvecs bvals dwi.mif with RealignTransform: 0 in the config file can lead to a mif file with incorrectly oriented gradient directions (e.g. on the HCP data).

Not sure how big a deal this is, but at least it was unexpected.

[Lestropie]

This will be the use case where my hypothetical in #2416 actually causes a demonstrable problem.

[diagiraldo]

Something similar is happening vice versa, when I have RealignTransform: 0 in .mrtrix.conf and convert from mif to nifti with -strides -1,2,3,4 the resulting gradients are not correct, here the example:

mrconvert DWI.mif -strides -1,2,3,4 -export_grad_fsl bvecs.txt bvals.txt DWI.nii.gz 
dwiextract DWI.nii.gz -singleshell - -fslgrad bvecs.txt bvals.txt | amp2sh - sh_DWI.mif -force 

I should say it doesn't happen with -strides 1,2,3,4:

mrconvert DWI.mif -strides 1,2,3,4 -export_grad_fsl bvecs.txt bvals.txt DWI.nii.gz 
dwiextract DWI.nii.gz -singleshell - -fslgrad bvecs.txt bvals.txt | amp2sh - sh_DWI.mif -force