no code implementations • 15 Feb 2024 • Jean-Marie Lemercier, Julius Richter, Simon Welker, Eloi Moliner, Vesa Välimäki, Timo Gerkmann
Here, we aim to show that diffusion models can combine the best of both worlds and offer the opportunity to design audio restoration algorithms with a good degree of interpretability and a remarkable performance in terms of sound quality.
1 code implementation • 18 Sep 2023 • Bunlong Lay, Jean-Marie Lemercier, Julius Richter, Timo Gerkmann
While the performance of usual generative diffusion algorithms drops dramatically when lowering the number of function evaluations (NFEs) to obtain single-step diffusion, we show that our proposed method keeps a steady performance and therefore largely outperforms the diffusion baseline in this setting and also generalizes better than its predictive counterpart.
1 code implementation • 22 Jun 2023 • Jean-Marie Lemercier, Joachim Thiemann, Raphael Koning, Timo Gerkmann
We show that our stochastic regeneration model outperforms other neural-network-based wind noise reduction methods as well as purely predictive and generative models, on a dataset using simulated and real-recorded wind noise.
1 code implementation • 21 Jun 2023 • Jean-Marie Lemercier, Simon Welker, Timo Gerkmann
We present in this paper an informed single-channel dereverberation method based on conditional generation with diffusion models.
no code implementations • 5 Jun 2023 • Danilo de Oliveira, Julius Richter, Jean-Marie Lemercier, Tal Peer, Timo Gerkmann
Since its inception, the field of deep speech enhancement has been dominated by predictive (discriminative) approaches, such as spectral mapping or masking.
no code implementations • 15 Mar 2023 • Julius Richter, Simon Welker, Jean-Marie Lemercier, Bunlong Lay, Tal Peer, Timo Gerkmann
In this paper, we present a causal speech signal improvement system that is designed to handle different types of distortions.
no code implementations • 1 Mar 2023 • Jean-Marie Lemercier, Julian Tobergte, Timo Gerkmann
We demonstrate that the resulting deep subband filtering scheme outperforms multiplicative masking for dereverberation, while leaving the denoising performance virtually the same.
2 code implementations • 22 Dec 2022 • Jean-Marie Lemercier, Julius Richter, Simon Welker, Timo Gerkmann
As diffusion models are generative approaches they may also produce vocalizing and breathing artifacts in adverse conditions.
1 code implementation • 4 Nov 2022 • Jean-Marie Lemercier, Julius Richter, Simon Welker, Timo Gerkmann
In this paper, we systematically compare the performance of generative diffusion models and discriminative approaches on different speech restoration tasks.
1 code implementation • IEEE/ACM Transactions on Audio, Speech, and Language Processing 2023 • Julius Richter, Simon Welker, Jean-Marie Lemercier, Bunlong Lay, Timo Gerkmann
This matches our forward process which moves from clean speech to noisy speech by including a drift term.
Ranked #19 on Speech Enhancement on VoiceBank + DEMAND
no code implementations • 6 Apr 2022 • Jean-Marie Lemercier, Joachim Thiemann, Raphael Koning, Timo Gerkmann
In this paper, a neural network-augmented algorithm for noise-robust online dereverberation with a Kalman filtering variant of the weighted prediction error (WPE) method is proposed.
no code implementations • 6 Apr 2022 • Jean-Marie Lemercier, Joachim Thiemann, Raphael Koning, Timo Gerkmann
This work focuses on online dereverberation for hearing devices using the weighted prediction error (WPE) algorithm.
no code implementations • 6 Apr 2022 • Jean-Marie Lemercier, Joachim Thiemann, Raphael Koning, Timo Gerkmann
By deriving new metrics analyzing the dereverberation performance in various time ranges, we confirm that directly optimizing for a criterion at the output of the multi-channel linear filtering stage results in a more efficient dereverberation as compared to placing the criterion at the output of the DNN to optimize the PSD estimation.