This repository contains clean templates and project examples to compile VST plugins with neural networks. A specific focus is given to deployment to embedded computers with the Elk Audio OS, a real-time Linux-based operating system for low latency embedded audio processing.
The repository is meant to accompany the following scientific paper:
Domenico Stefani, Luca Turchet "Real-Time Embedded Deep Learning on Elk Audio OS" in International Symposium on the Internet of Sounds (IS2), Pisa, Italy Oct 2023 (Accepted)
The paper will soon be available here
You can find the guide in the Wiki (Or the guide/README.md file).
This project is meant to be a guide to deploying neural networks in VST plugins, with a specific focus on Elk Audio OS.
Elk audio OS is a real-time Linux-based operating system for low latency embedded audio processing, which allows running VST plugins on several embedded computers.
This guide provides instructions for two different Inference Engines:
The folders in this repository are of two types: templates and examples.
The templates are meant to be used as a starting point for new projects, while the examples are meant to be used as a compilation and execution test.
Every subfolder provides configuration files to compile the project for both a regular Linux Laptop/Desktop machine (x86_64 architecture) and for the Elk Audio OS target (a Raspberry Pi 4 running Elk Audio OS, aarch64 architecture).
The repository contains the following folders:
ONNXruntime-VSTplugin-template/ONNXruntime-example/TFlite-VSTplugin-template/TFlite-example/
Domenico Stefani, 2023