If you finished a mix and it turns out that it serves the song a lot if you timestretch it -2bpm. What is the best way to do this, meaning;
1. Consolidate every track and timestretch each track individually OR
2. Timestretch it in the mastering. So, after the stereo mixdown, but before you apply masterbus compression, etc.? OR
3. A bit of both, print stems and timestretch those?
I would normally timestretch each track, but that has more to do with workflow in Ableton than anything else.
Had I been working in, as an example, Reaper I would probably stem it and stretch those, I've found that transient material like drums like other algorithms/settings than vocals or other tonally rich stuff.
Yeah, it's specifically the percussive instruments that bother me. For that matter, a piano or an acoustic guitar can be played percussively as well. That's why I thought that making stems will give the best result because, at least I think, that the more material you put in a track, the fewer spikes you end up with. No?