Implementation of deezer/spleeter using Vanilla PHP, recommended to separate songs instruments on "bulk" mode
Spleeter is Deezer source separation library with pretrained models written in Python and uses Tensorflow. It makes it easy to train source separation model (assuming you have a dataset of isolated sources), and provides already trained state of the art model for performing various flavour of separation.
This app was developed to create an easy way to separate instruments from audio files, using a web-based interface, turning it possible for non-technician users to create stems from batch files.
(If you need to persist your songs history, I recomend you take a look at https://github.com/ahmedash95/audio-spleeter)
You just need to have Docker installed on your computer
docker build . -t spleeter-php
docker run --name spleeter-php -p 81:80 --mount type=bind,source="$(pwd)",target=/var/www/html -d spleeter-php
- By default, php-spleeter will be looking for audio files over the
./audio-files
folder - Click the button to process the files
- Files processed will be on
./audio-files-processed
folder - Use interface controls to personalize the spleeter options
- Improve interface UX
- Add evaluate option on interface
- Create automated way to training AI models
The code of Spleeter is MIT-licensed.
I'll appreciate any kind of help \o/
- Deezer Research - Source Separation Engine Story - deezer.io blog post
- Music Source Separation tool with pre-trained models / ISMIR2019 extended abstract link
@article{spleeter2020,
doi = {10.21105/joss.02154},
url = {https://doi.org/10.21105/joss.02154},
year = {2020},
publisher = {The Open Journal},
volume = {5},
number = {50},
pages = {2154},
author = {Romain Hennequin and Anis Khlif and Felix Voituret and Manuel Moussallam},
title = {Spleeter: a fast and efficient music source separation tool with pre-trained models},
journal = {Journal of Open Source Software},
note = {Deezer Research}
}