Yep, the exact same problem arises in digital audio, mapping between integer sample formats and the floating point representation that is generally used internally.