With 32 bpc data, Apply Image subtract and add are not available. Apply Image subtract is the only way of taking the difference and retaining the negative values from the operation. This done by using and offset and scale which is available in the non 32bpc version. How might I simulate Apply Image without losing the negative values? Reducing to 16 bpc is not an option as I'm using 32 pbc for accuracy of an algorithm. Everything I try clips the negative values.
RONC