Rather than mess around trying all sorts of post-processing tricks on its line output signal to undo all the pre-processing that has been done inside the instrument to make its speakers sound good, why not try recording the digital as though it were an acoustic, i.e. with microphones?