# Data-Driven Room Acoustic Modeling Via Differentiable Feedback Delay Networks With Learnable Delay Lines

arXiv: https://arxiv.org/abs/2404.00082 · HTML: https://mlsanity.com/paper/2404.00082/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 50 references that are in the map.
- First posted on arXiv 29 Mar 2024.

## Builds on (most cited first, 8 of 8)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Automatic differentiation in machine learning: a survey](https://mlsanity.com/paper/1502.05767/) (2015)
- [DDSP: Differentiable Digital Signal Processing](https://mlsanity.com/paper/2001.04643/) (2020)
- [Cheap Orthogonal Constraints in Neural Networks: A Simple Parametrization of the Orthogonal and Unitary Group](https://mlsanity.com/paper/1901.08428/) (2019)
- [Differentiable Artificial Reverberation](https://mlsanity.com/paper/2105.13940/) (2021)
- [Efficient Synthesis of Room Acoustics via Scattering Delay Networks](https://mlsanity.com/paper/1502.05751/) (2015)
- [On Lossless Feedback Delay Networks](https://mlsanity.com/paper/1606.07729/) (2016)
- [Accelerating Parameter Extraction of Power MOSFET Models Using Automatic Differentiation](https://mlsanity.com/paper/2110.15048/) (2021)

## Built on by (most cited first, 1 of 1)

- [Sines, Transient, Noise Neural Modeling of Piano Notes](https://mlsanity.com/paper/2409.06513/) (2024)

## Abstract (arXiv)

Over the past few decades, extensive research has been devoted to the design of artificial reverberation algorithms aimed at emulating the room acoustics of physical environments. Despite significant advancements, automatic parameter tuning of delay-network models remains an open challenge. We introduce a novel method for finding the parameters of a Feedback Delay Network (FDN) such that its output renders target attributes of a measured room impulse response. The proposed approach involves the implementation of a differentiable FDN with trainable delay lines, which, for the first time, allows us to simultaneously learn each and every delay-network parameter via backpropagation. The iterative optimization process seeks to minimize a perceptually-motivated time-domain loss function incorporating differentiable terms accounting for energy decay and echo density. Through experimental validation, we show that the proposed method yields time-invariant frequency-independent FDNs capable of closely matching the desired acoustical characteristics, and outperforms existing methods based on genetic algorithms and analytical FDN design.