Differentiable Artificial Reverberation
Preprint
Abstract
Artificial reverberation (AR) models play a central role in various audio applications. Therefore, estimating the AR model parameters (ARPs) of a target reverberation is a crucial task. Although a few recent deep-learning-based approaches have shown promising performance, their non-end-to-end training scheme prevents them from fully exploiting the potential of deep neural networks. This motivates to introduce differentiable artificial reverberation (DAR) models which allows loss gradients to be back-propagated end-to-end. However, implementing the AR models with their difference equations "as is" in the deep-learning framework severely bottlenecks the training speed when executed with a parallel processor like GPU due to their infinite impulse response (IIR) components. We tackle this problem by replacing the IIR filters with finite impulse response (FIR) approximations with the frequency-sampling method (FSM). Using the FSM, we implement three DAR models -- differentiable Filtered Velvet Noise (FVN), Advanced Filtered Velvet Noise (AFVN), and Feedback Delay Network (FDN). For each AR model, we train its ARP estimation networks for analysis-synthesis (RIR-to-ARP) and blind estimation (reverberant-speech-to-ARP) task in an end-to-end manner with its DAR model counterpart. Experiment results show that the proposed method achieves consistent performance improvement over the non-end-to-end approaches in both objective metrics and subjective listening test results.Keywords
All Related Versions
This publication has 26 references indexed in Scilit:
- Lightweight and Interpretable Neural Modeling of an Audio Distortion Effect Using Hyperconditioned Differentiable Biquads2021
- Pyroomacoustics: A Python Package for Audio Room Simulation and Array Processing Algorithms2018
- Statistics of natural reverberation enable perceptual separation of sound and spaceProceedings of the National Academy of Sciences, 2016
- Estimation of Room Acoustic Parameters: The ACE ChallengeIEEE/ACM Transactions on Audio, Speech, and Language Processing, 2016
- Efficient Synthesis of Room Acoustics via Scattering Delay NetworksIEEE/ACM Transactions on Audio, Speech, and Language Processing, 2015
- A Perceptual Study on Velvet Noise and Its Variants at Different Pulse DensitiesIEEE Transactions on Audio, Speech, and Language Processing, 2013
- Fifty Years of Artificial ReverberationIEEE Transactions on Audio, Speech, and Language Processing, 2012
- A just noticeable difference in C 50 for speechApplied Acoustics, 1999
- Circulant and elliptic feedback delay networks for artificial reverberationIEEE Transactions on Speech and Audio Processing, 1997
- Identification of systems containing linear dynamic and static nonlinear elementsAutomatica, 1982