Sometimes I see this warning message when a dset is read into AFNI – here, into 3dinfo (and this dataset came from some other software output):
** AFNI converts NIFTI_datatype=2 (UINT8) in file /home/sweethome/some/path/some_dset.nii.gz to FLOAT32
Warnings of this type will be muted for this session.
Set AFNI_NIFTI_TYPE_WARN to YES to see them all, NO to see none.
How much should I care or be concerned about the above warning?
It clogged my happy data viewing, so I ran “3dinfo -DAFNI_NIFTI_TYPE_WARN=NO …” to get rid of it temporarily. Can I feel secure enough that I am not missing anything vital to put:
Based on my reading of thd_niftiread.c, that warning should appear only in cases of data formats that are not used directly in AFNI, such as the upscaling of unsigned 16-bit to floating point format, and in the case of NIFTI data with a scale factor. In neither case typically,this should not be an important issue. The first case results in a dataset that’s double the expected size in memory, but disk space with compression is likely to be similar. The second case indicates a difference in the voxel data to reflect the scaling from the scl_slope and scl_intercept in the NIFTI header; in which case, the data is also converted to floating point format. That up-conversion can potentially occur for 8-bit byte or 16-bit integer data, even if otherwise a known AFNI format. I haven’t seen this problem myself, but it is potentially there. You can happily put the blindfold over your eyes for most purposes with that environment variable.
… so indeed there is a “slope” being applied, at least.
thanks,
pt
The
National Institute of Mental Health (NIMH) is part of the National Institutes of
Health (NIH), a component of the U.S. Department of Health and Human
Services.