This python based software uses ultralytic's YoloV8 (You Only Look Once) state-of-the-art video recognition deep learning model to identify, count and track the number of people on a webcam. It asignes one granular synthetizer to each person and, depending on its X-Y position, the number of people, and the closeness between them, some parameters of the synthetizers - such as filter, pitch shift, panning and reverb - are modified in real-time.
This project was built for the Chilean 2023 exhibition 'Aparatos de Observación Pluvial', founded by the Chilean Ministry of Culture, Art and Patrimony, FONDART 2022. Folio: 670933.
CONDA_SUBDIR=osx-64 conda create -n yolo python=3.8 wheel=0.38.4 ffmpeg=4.2.2 portaudio=19.6.0 pyinstaller=5.6.2
conda activate yolo
conda config --env --set subdir osx-64
pip install -r requirements.txt
conda create -n yolo python=3.8
conda activate yolo
pip install -r requirements.txt
download audios from google drive, uncompress it and place it on root directory.
python aparalius.py
conda activate yolo
git clone git@github.com:ultralytics/ultralytics.git
pyinstaller .\aparalius.spec --noconfirm
mkdir dist/audios
mkdir dist/audios/Final
mkdir dist/audios/IRs
mv your-audios.wav dist/audios/Final/your-audio.wav
mv your-IR.wav dist/audios/Final/IRs/your-IR.wav