-
15 Feb 2024 - Diffusion Models for Audio Restoration
Submitted to IEEE Signal Processing Magazine.
- Keywords :
-
- Diffusion Models
- Overview
- Audio Restoration
- Model-based Signal Processing
-
30 Oct 2023 - Diffusion posterior sampling for informed single-channel dereverberation
Accepted at WASPAA 2023.
- Keywords :
-
- Diffusion Models
- Dereverberation
- Inverse Problems
- Posterior Sampling
-
15 Sep 2023 - Wind Noise Reduction with a Diffusion-based Stochastic Regeneration Model
Accepted to VDE 15th ITG conference on Speech Communication.
- Keywords :
-
- Diffusion Models
- Wind Noise
- Speech Enhancement
-
01 Aug 2023 - StoRM - A diffusion-based stochastic regeneration model for speech enhancement and dereverberation
Published in IEEE Transactions on Audio, Speech and Language Processing.
- Keywords :
-
- Diffusion Models
- Speech Enhancement
- Dereverberation
- Neural network
-
30 May 2023 - Extending DNN-based Multiplicative Masking to Deep Subband Filtering for Improved Dereverberation
Accepted at Interspeech 2023.
- Keywords :
-
- Multi-frame filtering
- Subband approximation
- Dereverberation
- Denoising
- Neural network
-
01 Apr 2023 - A Neural Network-supported Two-stage Algorithm for Lightweight Dereverberation on Hearing Devices
Accepted at EURASIP Journal on Audio, Speech, and Music Processing.
- Keywords :
-
- Dereverberation
- Neural network
- End‑to‑end learning
- Hearing devices
-
01 Dec 2022 - Analysing Diffusion-based Generative Approaches versus Discriminative Approaches for Speech Restoration
Accepted at ICASSP 2023.
- Keywords :
-
- Generative modelling
- Diffusion models
- Speech enhancement
- Dereverberation
- Bandwidth extension
-
01 Jun 2022 - Neural Network-augmented Kalman Filtering for Robust Online Speech Dereverberation in Noisy Reverberant Environments
Accepted at Interspeech 2022.
- Keywords :
-
- Dereverberation
- Kalman Filtering
- Adaptive Processing
- Neural Network
- End-to-end Training
-
01 Nov 2021 - Customizable end-to-end optimization of online neural network-supported dereverberation for hearing devices
Accepted at ICASSP 2022.
- Keywords :
-
- Online Algorithm
- Dereverberation
- Neural Network
- End-to-end Learning
- Hearing devices
-
01 Sep 2021 - An Integrated Deep Clustering-Based System for Speaker Count Agnostic Speech Separation
Accepted at 14th ITG Conference on Speech Communication.
- Keywords :
-
- Speech Separation
- Speaker Counting
- Deep Clustering
- Multi-task Learning
-
06 Feb 2020 - Bio-inspired Encoders for Temporal Convolutional Networks performing Speech Enhancement and Separation
Imperial College London MSc Thesis supervised by Dr Dan F. M. Goodman and Prof. Dr-Ing. Timo Gerkmann on bio-inspired neural networks for speech enhancement and separation.
- Keywords :
-
- Auditory modelling
- Neural Networks
- Nonlinear Filters
- PyTorch
- Bio-inspired technologies
-
13 Jan 2020 - Electroacoustic study of new types of loudspeakers
A 6-month IRCAM-led study on innovative loudspeaker prototypes by Voxline Audio. Traditional measurement methods were used as well as innovative Post-Hamiltonian Systems modelling and new harmonic content analysis.
- Keywords :
-
- HiFi
- Loudspeaker
- Acoustics
- Nonlinear multiphysics modelling
- Port-Hamiltonian Systems
- Homophase Signals