Definitely separate tracks for me, but I’m definitely still experimenting and searching.
I did experiment with several things over the years, from creating everything as MIDI tracks first, and then recording one thing at a time (I didn’t have a multi-channel interface) to just recording the stereo out.
Two things seem to somehow work for me, depending on the type of track I’m making:
- Either treat everything as if it was a field recording (I have to thank Matthias Puech for opening my eyes on this!) or
- play the whole thing as if it was a liveset (after rehearsing it for a bit) and then do a multi-track recording (thankfully I now have an interface with enough inputs)
The first one seems to work better for my normal practice, because I think that I tend to compose in the DAW by cut, pasting and mangling audio. I do use a lot of field recordings, so I guess I tend to apply the same logic to synthesized sound as well. I often come up with a patch consisting of just one voice, record that and store it on my HD. Later I’ll maybe combine that with some other sounds to make a track.
The downside of this approach is that for me it does not work well for more structured composition. That’s where the multi-track approach is a lot better.
Now, the challenge is probably to find a way to combine the two somehow. That’s where I’m kinda stuck right now.