afni.nimh.nih.gov

Also, the latest version of 3dcalc should not set a scale factor with such a calculation. First, the outputs are calculated in floating point. It then checks the output values, and if they are all integers and all fit into the bit length of the datum (16 bits for short, 8 bits for byte), then no scaling will be done. However, this change is fairly recent, so you may not be up-to-date.

bob cox