Sample application connecting to A2F Controller

A sample application is provided to demonstrate how to communicate with the Audio2Face microservices. This python application will interact with A2F controller.

Assumptions

The Audio2Face and Audio2Face Controller Microservices are up and running.

Sending an audio clip

Clone the repository: https://github.com/NVIDIA/ACE

Follow the instructions in the above link to set up the specified environments.

Go to microservices/audio_2_face_microservice/scripts/audio2face_microservices_interaction_app subfolder.

And follow the setup instructions in the README.md.

This sample python application can be used as follows:

$ ./a2f.py --help
$ ./a2f.py <audio_file.wav> <config.yml> -u <ip>:<port>

For example,

$ ./a2f.py audio.wav config_mark_v2.yml -u 127.0.0.1:52000

  • The script requires two parameters: an audio file in PCM 16-bit format and a YAML configuration file containing emotion parameters.

  • Additionally, it accepts a -u parameter for the A2F Controller URL. For quick start deployment, use 127.0.0.1:52000.

  • To test the script, you’ll need to provide an audio file.

Results

This will produce a folder in which 5 files are available:

$ ls -l 20240410_120000
-rw-rw-r-- 1 user user   1733 Apr 10 12:00 a2e_emotion_output.csv
-rw-rw-r-- 1 user user    466 Apr 10 12:00 a2f_input_emotions.csv
-rw-rw-r-- 1 user user  17702 Apr 10 12:00 a2f_smoothed_emotion_output.csv
-rw-rw-r-- 1 user user  76219 Apr 10 12:00 animation_frames.csv
-rw-rw-r-- 1 user user 133368 Apr 10 12:00 out.wav

  • out.wav: contains the audio received

  • animation_frames.csv: contains the blendshapes

  • a2f_input_emotions.csv: contains the emotions provided as input in the gRPC protocol

  • a2e_emotion_output.csv: contains a blend of a2f_input_emotions and emotions inferred by Audio2Emotion

  • a2f_smoothed_emotion_output.csv: contains the emotions of a2e_emotion_output smoothed over time