Once the todos below have been implemented, the Mienophone VCV plugin will:
- capture pictures from a webcam
- trade the pictures for emotion parameters using MS Face API (internet connection required)
- map emotion parameters onto module outputs.
- API requests must be non-blocking (Multithreading / Async / curl "hiper")
- webcam capture (we could run this separately, e.g.
fswebcam -q -r 1280x720 --crop 480x720,500x0 capture.jpg
) - add clock/gate input
- fix memory leaks
- cleanup the ugliness from shameless green code.
- cross-plattform (mac/bsd)
- test if this benefits from video streams instead of posting images
- build-essential
- curl API
- v4l - see example code but this is not cross-plattform compatible
- VCV Rack SDK
apt install libcurl4-openssl-dev build-essential
git clone https://github.com/stellare-modular/vcv-rack-sdk.git
echo "export RACK_DIR=<Rack SDK folder>" >> ~/.bashrc
See also VCV Rack PluginDevelopmentTutorial
export RACK_DIR=$HOME/Projects/vcvrack/Rack-SDK
make
make dist
make install
docker build . -t fake-api
docker run -it --rm -d --name fake-api -p 8080:8080 fake-api
curl -i http://localhost:8080/face/v1.0/detect