"Transparent zero" operation for 16/32-bit images?

Posted by Mike Cowperthwaite on Oct 07, 2009; 8:28pm
URL: http://imagej.273.s1.nabble.com/Transparent-zero-operation-for-16-32-bit-images-tp3690873.html

I've been using "Transparent zero" both from the Image Calculator and
pasting, allowing me to quickly mask out parts of an image.  (That is,
the second operator of the calculator, or the image being pasted, is a
mask with zeros where I want to keep the image, and nonzero where I want
to force the value to a fixed level; I've been using a masking pixel
value of "1" to make those areas appear black.)

Previously, I'd been using 8-bit greyscale.  Now I want to do the same
operation on a 16-bit image, and my efforts are not working -- the
results always seems to be simple copy of the mask image.  I can convert
the 16-bit to 8-bit, but then I'm losing resolution (the masked-out
portions are the ones that show brightest, and the purpose is to
eliminate those bright features while maintaining as much detail as
possible with what's left over.)

Is this something that works, but I'm doing it wrong, or something that
just doesn't work?  If the latter, can it be made to work?

--
Mike Cowperthwaite
Lathrop Engineering, San Jose CA