I tested if Audio Driven Animation for MetaHuman Animator works with a song. I separated lead vocal from the song and generated facial animation from the lead vocal audio using Audio Driven Animation for MetaHuman Animator. (UE5.5.4 Experimental)
The MetaHuman looked a bit angry with the original generated animation so I mixed City Sample Crowds Crowd_Happy animation and slightly modified manually with Face Control Rig.
The mouth doesn’t open enough because I mixed City Sample Crowds animation that mouth is closed but overall it looks better than I was thinking.