GRAFX: An Open-Source Library for Audio Processing Graphs in PyTorch

TL;DR

GRAFX is an open-source PyTorch library that enables efficient construction and optimization of audio processing graphs, demonstrated through a music mixing scenario with GPU acceleration.

Contribution

It introduces GRAFX, a novel library that facilitates parallel GPU computation of audio graphs and enables gradient-based optimization of audio processing parameters.

Findings

Efficient parallel GPU computation of audio graphs.

Successful application in music mixing with gradient descent optimization.

Open-source availability for community use.

Abstract

We present GRAFX, an open-source library designed for handling audio processing graphs in PyTorch. Along with various library functionalities, we describe technical details on the efficient parallel computation of input graphs, signals, and processor parameters in GPU. Then, we show its example use under a music mixing scenario, where parameters of every differentiable processor in a large graph are optimized via gradient descent. The code is available at https://github.com/sh-lee97/grafx.