FLAMO: An Open-Source Library for Frequency-Domain Differentiable Audio Processing

TL;DR

FLAMO is an open-source library that enables the design and optimization of differentiable linear time-invariant audio systems using frequency-sampling methods, facilitating integration with neural networks.

Contribution

It introduces a flexible, easy-to-use framework for creating differentiable audio modules based on frequency-sampling filter design, with practical case studies demonstrating its effectiveness.

Findings

Successful optimization of an artificial reverberator.

Improved response coloration in active acoustics system.

Accessible interface for constructing and training audio modules.

Abstract

We present FLAMO, a Frequency-sampling Library for Audio-Module Optimization designed to implement and optimize differentiable linear time-invariant audio systems. The library is open-source and built on the frequency-sampling filter design method, allowing for the creation of differentiable modules that can be used stand-alone or within the computation graph of neural networks, simplifying the development of differentiable audio systems. It includes predefined filtering modules and auxiliary classes for constructing, training, and logging the optimized systems, all accessible through an intuitive interface. Practical application of these modules is demonstrated through two case studies: the optimization of an artificial reverberator and an active acoustics system for improved response coloration.