Sure, that's one way to do it, but if the invalid byte were logically tied to the byte before it in its original encoding, then now your substring split will mangle that character.