The whole point of 24-bit is to give you flexibility when the singer decides that he's going to whisper one verse from three feet away and then shove the mic down his throat while he screams the chorus. Or you can be sloppy when you set the preamp gain—just give yourself enough headroom, and you'll adjust the levels later. Notice that I said "later", not "never". Even the most amateurish home engineer with a hangover the size of Texas is going to adjust the levels of different tracks in the mix. Once you've done the mixdown, the levels are "reasonable", and you can listen at home on 16-bit system. You won't hear the difference.
No mastering, no compression. 16-bit audio is still enough, once you've mixed a song.