Last year there was a question regarding converting AFNI sets to ANALYZE shorts for use with AIR.
[
afni.nimh.nih.gov]
Rick suggested using 3dcalc with the -nscale, with the caveat: "you are not worried about truncation errors, and the maximum values in the dataset do not exceed the maximum values for the integer types"
With this in mind, does it make sense to run 3dcalc first with the -gscale the run 3dcalc on the resulting dataset with the -nscale? Does this insure that the data will be in the right range?
ex:
3dcalc -datum short -gscale -a fmri+orig -expr a -prefix fmri_gshort
3dcalc -datum short -nscale -a fmri_gshort+orig -expr a -prefix fmri_short
3dAFNItoANALYZE -4D fmri_short fmri_short+orig
The results "look" okay so far, but am I missing something?