Casting 0xFFFFFFFF to float does not give NaN as is needed here. Found through compiler warning which is now fixed.
Files: logImageLib.h, logImageLib.c: combined cineon/dpx image library dpxlib.h, dpxlib.c: dpx specific library dpxfile.h: dpx file structure cineonlib.h, cineonlib.c: cineon specific library cineonfile.h: cineon file structure logImageCore.h, logImageCore.c: log image routines common to cineon amd dpx