What you’re describing is actually standard practice in most feature dubbing stages. What we normally do is prepare everything at 100% speed (24fps @ 48k) but then pull the WHOLE mix stage down to 29.97 / 23.98 / 47.952k. That is, we pull the audio AND the video AND the timecode down by that old .1% NTSC pulldown. This effectively keeps all the elements in perfect sync, and allows us to use mixing consoles, outboard gear, and video equipment that only operate in the 29.97 TC world. I realize it seems a little archaic and overly complicated, but it’s actually very straightforward in practice. Once the mix is finished, the final master will be in perfect sync when you play it back at 48k with 24fps picture (AKA film).