Drop new audio (mp3/wav) here, then register it for the viewer with:

  cd /home/ubuntu/service/kemix-engine/package/face/animasync-face-v3
  /data/miniconda3/bin/python tools/export/add_audio.py --audio data/audio_new/<file> --sid <id> --emotion <label> --text "대사"

emotion labels: neutral joy laughter excitement agreement gratitude sadness crying sulk apology struggle anger refusal surprise fluster shy
This folder lives on the SERVED tree (/home/ubuntu/service/...); /home/seungeun/kemix-engine/.../data/audio_new is a symlink to it.