I'd like to isolate the vocals on a track I have using Logic. Can anyone provide some advice on how to do this please?
You could try also this tricky method. No miracle, just tricky, but it's time-saving because performs a sort of auto tuning. Just saying, it's the opposite of cancelling a vocal from a stereo track.
Provided that most solo vocals are mixed "in front" (no panning), when you process the stereo track in a Mid/Side network (M=L+R and S=L-R) and discard the M(id) information, the remaining S(ide) channel has a good attenuation on the direct signal of every sound mixed in center.
I say "direct" (=everithing is exactly in phase) because FXs do not behave this way. For this, a very wet signal (=heavily processed) will be hard to isolate. For this reason, this method has very little effect on live takes made with field recorders.
Bounce the S channel on a new track (and mute the original), plug in the MatchEQ and let it learn its frequency response. Flip the curve (save it) and apply it on the original track. With MatchEQ you can also tweak the resolution of the curve and the overall percentage.
Warning! Mangle ONLY duplicate files!
In most cases the M channel is useless, because good mixing engineers always check mono compatibility of their works, but it might be convenient to try MatchEQ on it, as a second chance.