I don't disagree. It's temporary, and just meant to be an arrangement demo, of course. Julia will ultimately have time to record, but in the meantime, this will serve better (I think) as a representation of what the arrangement does than my old routine of just pitching my vocals up. When Julia sings, I won't be using any Melodyne, and it will be more optimal and soulful. But yes, go humans!

Hey Jake...yep, it's super dicey and we're entering a new era to be sure. Struggled with whether or not to use the models for this purpose, but ulimately made the call that being able to offer directors more options for programming is better than letting the charts just sit on my hard drive unsung by the actual humans in the world who will bring these to life for real. As another commenter noted, it's pretty rough getting AI models to have soul (and they rely on the soulfulness of the input vocalist, which in this case was me, and I'm certainly lacking in that area).