I think if you shared what you are trying to do, then it would help others understand.
One advantage of floating point values is that you can adjust the values beyond the clipping ranges and recover them. This is helpful for adjusting things like exposure in photos.
Png files can also be higher bit depth.
Also, I think GEGL works in higher bit depth, as well as GMic. So the filters will be more accurate with less artifacts.
--edit. I did a test, saving a 16 bit tif file from Blender. Gimp 2.10 opened it as 16bit
One advantage of floating point values is that you can adjust the values beyond the clipping ranges and recover them. This is helpful for adjusting things like exposure in photos.
Png files can also be higher bit depth.
Also, I think GEGL works in higher bit depth, as well as GMic. So the filters will be more accurate with less artifacts.
--edit. I did a test, saving a 16 bit tif file from Blender. Gimp 2.10 opened it as 16bit