Ran a real voice track through an anime-style character and let the audio drive the whole performance, not a fixed timing template. The mouth shapes, the pauses, even the small head tilts came straight from the waveform. Curious how many of you are doing audio-driven animation yet versus hand-keying visemes.