So that I can tune it, I am trying to remove a lead vocal from the audience tracks in a live performance. What I envision is being able to insert a key from the lead vocal track, keyed into an aux track with the audience tracks, and having Clarity remove exactly and only the lead vocal. The audience is singing along, which I want to retain of course, but I need to get rid of the lead.
I imagine that this would be a very desirable feature for many people!
Actually, I thought if you rotated the control all the way to the left it actually removes the voice.
Failing that, the other thing you could try is create a parallel track, use Clarity to remove the music, then flip the phase on the track and blend it with the original so it is phase cancelling the vocal.
I’m not sure this will be 100% effective, but worth an experiment.
You’re right, moving the dial to the left instead of right removes the vocal from a track and leaves the ambience behind, you don’t even need to do the phase reversal method.
But what I am talking about is: removing a lead vocal picked up by ambience mics that are elsewhere in a venue (I’m working on a live recording).
I can put Clarity onto the audience tracks, and it does a good job of removing the lead vocal, but it also takes away a fair bit of the crowd singing along, chanting, cheering, and so on. The stuff that makes a live recording great!
Clarity was programmed using machine learning to to be able to separate out vocals in general. I’m hoping that it can machine-learn a particular vocal performance, as opposed to sort of a generic filter.
What I am asking for is: the ability for Clarity to generate a “mask” (like in Photoshop, if you’re familiar) from that exact lead vocal, which could be then applied via Clarity inserted on other tracks, to remove precisely, and only, that one specific vocal.