I take this as a suggestion to implement a function for reading the relative instead of the absolute bone rotation.