/soundsride

SoundsRide: Affordance-Synchronized Music Mixing for In-Car Audio Augmented Reality [Server]

Primary LanguagePython

SoundsRide: Affordance-Synchronized Music Mixing for In-Car Audio Augmented Reality [Server]

Mohamed Kari1,2, Tobias Grosse-Puppendahl1, Alexander Jagaciak1, David Bethge1, Reinhard Schütte2, Christian Holz3

1 Porsche AG, Stuttgart, Germany
2 University of Duisburg-Essen, Germany
3 Sensing, Interaction & Perception Lab, Department of Computer Science, ETH Zürich, Switzerland


SoundsRide runs the mixing server on MacOS and the localization client on iOS. This repo contains the mixing server. For the localization client, refer to the iOS repo.

Preparations

Install dependencies:

brew install portaudio
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Place the following songs into the ./tests/data/ directory:

  • animals.mp3 (Martin Garrix, 05:02)
  • drink-up-me-hearties.mp3 (Hans Zimmer, 04:29)
  • requiem-for-a-tower.mp3 (Clint Mansell, 04:10)
  • music.mp3 (John Miles, 05:52)
  • river-flows-in-you.mp3 (Yiruma, 03:29)
  • tsunami.mp3 (DVBSS & Borgeous, 03:57)
  • shot-me-down.mp3 (David Guetta ft. Skylar Grey, 03:09)
  • underground.mp3 (Lindsey Stirling, 04:25)

These mp3 along with the transition specs will be loaded into the song snippet database.

Try out basic component functionalities

try out the hot-swappable player

python -m soundsride.player $PATH_TO_MP3 
# continuous playback

try playback visualization

python -m soundsride.viz_player $PATH_TO_MP3
# starting playback, output `latest_audio.jpg` in the current folder and continuoulsy overwrite it

try the transition visualizer

python -m soundsride.canvas.transition_spec_canvas $PATH_TO_MP3
# outputs `transitiion_*.jpg` into the current folder, indicating the transition markers

Build the service proto buffer code and spin up the server

Communication is implemented via proto-buffer-based gRPC.

make python-protos
python -m soundsride.service.server 8888

As soon as the client starts sending localization data, the server will mix music accordingly, write the received localization data as well as the system state overview visualizations to the log/$TIMESTAMP directory.

real-time mixing