This brings up an interesting question I've not pondered before: just how close together do automation nodes
need to be?
You're assuming that finer resolution would result in smoother-sounding changes. I'm sure that's true, but obviously only up to a point.
Decreasing the interval between automation nodes down to 2 probably wouldn't help anyway. I'm guessing the fastest change you could actually
hear would be around 10ms, and even then it would have to be a drastic jump - and that's assuming the effect/synth you're automating can even respond that quickly. We're talking an interval far less than the duration of a single sung consonant.
I'm going to try to find some information...