audio-11

Speech Dereverberation

3260 papers • 126 benchmarks • 313 datasets

Removing reverberation from audio signals

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in speech-dereverberation-23

Trend

Dataset

Best Model

Actions

WHAMR!

Deep Noise Suppression (DNS) Challenge

WHAMR_ext

Libraries

i

Use these libraries to find speech-dereverberation-23 models and implementations

Datasets

Subtasks

No subtasks available.

Most implemented papers

Deformable Temporal Convolutional Networks for Monaural Noisy Reverberant Speech Separation

William Ravenscroft, Stefan Goetze, Thomas Hain•Wed Oct 26 2022

Deletable convolution is proposed as a solution to allow temporal convolutional networks models to have dynamic RFs that can adapt to various reverberation times for reverberant speech separation.

13

Content

WHAMR_ext

spatialized WSJCAM0

0

Paper Graph

MESH2IR: Neural Acoustic Impulse Response Generator for Complex 3D Scenes

Dinesh Manocha, Anton Ratnarajah, Zhenyu Tang, R. Aralikatti•Tue May 17 2022

This work proposes a mesh-based neural network (MESH2IR) to generate acoustic impulse responses (IRs) for indoor 3D scenes represented using a mesh and is the first neural-network-based approach to predict IRs from a given 3D scene mesh in real-time.

50 0

Paper Graph

StoRM: A Diffusion-Based Stochastic Regeneration Model for Speech Enhancement and Dereverberation

Julius Richter, Simon Welker, Jean-Marie Lemercier, Timo Gerkmann•Wed Dec 21 2022

This work presents a stochastic regeneration approach where an estimate given by a predictive model is provided as a guide for further diffusion, and shows that this approach enables to use lighter sampling schemes with fewer diffusion steps without sacrificing quality, thus lifting the computational burden by an order of magnitude.

168 0

Paper Graph

Investigating Generative Adversarial Networks based Speech Dereverberation for Robust Speech Recognition

Junbo Zhang, Yujun Wang, Lei Xie, Ke Wang, Sining Sun, Fei Xiang•Mon Mar 26 2018

Deep investigations in the use of GAN-based dereverberation front-end in ASR find that LSTM leads a significant improvement as compared with feed-forward DNN and CNN in the dataset and it is important to update the generator and the discriminator using the same mini-batch data during training.

39 0

Paper Graph

HiFi-GAN: High-Fidelity Denoising and Dereverberation Based on Speech Deep Features in Adversarial Networks

Adam Finkelstein, Jiaqi Su, Zeyu Jin•Tue Jun 09 2020

This paper introduces HiFi-GAN, a deep learning method to transform recorded speech to sound as though it had been recorded in a studio, trained with multi-scale adversarial discriminators in both the time domain and the time-frequency domain.

155 0

Paper Graph

Late reverberation suppression using U-nets

D. Le'on, Felipe A. Tobar•Mon Oct 04 2021

A supervised dereverberation technique using fully-convolutional encoder-decoder networks with layers arranged in the form of an U and connections that skip some layers is proposed, which improves quality, intelligibility and other performance metrics compared to the original U-net method and it is on par with the state-of-the-art GAN-based approaches.

5 0

Paper Graph

Task-Specific Optimization of Virtual Channel Linear Prediction-Based Speech Dereverberation Front-End for Far-Field Speaker Verification

Joon-Young Yang, Joon‐Hyuk Chang•Sun Dec 26 2021

A novel front-end design that involves a recently proposed extension of the weighted prediction error (WPE) speech dereverberation algorithm, the virtual acoustic channel expansion (VACE)-WPE, demonstrating its superiority over fully neural front-ends and other TSO methods in various cases.

3 0

Paper Graph

Receptive Field Analysis of Temporal Convolutional Networks for Monaural Speech Dereverberation

William Ravenscroft, Stefan Goetze, Thomas Hain•Tue Apr 12 2022

Analysis of dereverberation performance depending on the model size and the RF of Temporal convolutional net-works shows that a larger RF can have significant improvement in performance when training smaller TCN models.

9 0

Paper Graph

Adding a benchmark result helps the community track progress.