A reference image, an audio track, and a prompt become a lip-synced video. Humans or not, speech or song, close portrait or wider frame, steerable mid-take and live when you need it — that's rave.
Speech in, face out — mouth, breath, emphasis all in sync.
Speech works. Song does too. The face matches either.
Swap references live. The face keeps speaking, in a new style.
A wider frame: shoulders move, hands gesture, the body plays along.
Fast enough for a live exchange. You talk; it talks back.
some ravers may be missing — no avatar in slack, or a failed generation ;(