This Python script leverages the headless mode of Audio2Face to generate lip and face animations for characters. It provides methods to control the Audio2Face headless server and interact with it through a requests API.

A use case is to generate animations for a batch of audio files and export them as USD files for example for Maya or Unreal Engine 5.

Big thank you for the NVIDIA Team who made Audio2Face a great tool.

• Python 3.x
• Audio2Face installed on the system

• Modify the ROOT_DIR and DEFAULT_OUTPUT_DIR variables in the settings.py file as needed.

Feel free to explore and customize the script based on your project requirements. Enjoy animating your characters with Audio2Face!

Install the py_audio2face package using pip:

pip install py_audio2face
or 
pip install git+https://github.com/w4hns1nn/py_audio2face.git

1. Initialize Audio2Face Instance:

import py_audio2face as pya2f
a2f = pya2f.Audio2Face()

2. Generate Animation for a Single Audio File:

audio_file_path = "path/to/audio/file.wav"
output_path = "path/to/output/animation.usd"
a2f.audio2face_single(audio_file_path, output_path, fps=60)

3. Generate Animations for an Entire Folder:

input_folder = "path/to/audio/folder"
output_folder = "path/to/output/folder"
a2f.audio2face_folder(input_folder, output_folder)

4. Shutdown Audio2Face Server:

a2f.shutdown_a2f()

This example initializes an Audio2Face instance, processes audio files in a specified folder, and shuts down the Audio2Face server.

Any contribution is appretiatet. At the moment the tests are not working and should be added. Also the communication with the headless.bat file and shutting down the server can be improved. Files with presettings for different characters would be great.

Me the author plans to use the package in another package that handles audio generation with tts and bark and transforms them instantly with a2f.