English
Related papers

Related papers: An Analysis of the Variance of Diffusion-based Spe…

200 papers

Diffusion models have emerged as the de facto choice for generating high-quality visual signals across various domains. However, training a single model to predict noise across various levels poses significant challenges, necessitating…

Computer Vision and Pattern Recognition · Computer Science 2024-11-28 Tiankai Hang , Shuyang Gu , Xin Geng , Baining Guo

We explore unsupervised speech enhancement using diffusion models as expressive generative priors for clean speech. Existing approaches guide the reverse diffusion process using noisy speech through an approximate, noise-perturbed…

Sound · Computer Science 2025-07-04 Mostafa Sadeghi , Jean-Eudes Ayilo , Romain Serizel , Xavier Alameda-Pineda

Denoising diffusion probabilistic models have been recently proposed to generate high-quality samples by estimating the gradient of the data density. The framework defines the prior noise as a standard Gaussian distribution, whereas the…

Machine Learning · Statistics 2022-02-22 Sang-gil Lee , Heeseung Kim , Chaehun Shin , Xu Tan , Chang Liu , Qi Meng , Tao Qin , Wei Chen , Sungroh Yoon , Tie-Yan Liu

Generative speech enhancement has recently shown promising advancements in improving speech quality in noisy environments. Multiple diffusion-based frameworks exist, each employing distinct training objectives and learning techniques. This…

Audio and Speech Processing · Electrical Eng. & Systems 2025-01-22 Julius Richter , Danilo de Oliveira , Timo Gerkmann

In this study, we aim to explore the effect of pre-trained conditional generative speech models for the first time on dysarthric speech due to Parkinson's disease recorded in an ideal/non-noisy condition. Considering one category of…

Audio and Speech Processing · Electrical Eng. & Systems 2024-12-19 Joanna Reszka , Parvaneh Janbakhshi , Tilak Purohit , Sadegh Mohammadi

By learning the gradient of smoothed data distributions, diffusion models can iteratively generate samples from complex distributions. The learned score function enables their generalization capabilities, but how the learned score relates…

Machine Learning · Computer Science 2024-12-16 Binxu Wang , John J. Vastola

Speech enhancement (SE) performance is known to depend on noise characteristics and signal to noise ratio (SNR), yet intrinsic properties of the clean speech signal itself remain an underexplored factor. In this work, we systematically…

Audio and Speech Processing · Electrical Eng. & Systems 2025-09-24 Mingchi Hou , Ina Kodrasi

Speech enhancement is an essential task of improving speech quality in noise scenario. Several state-of-the-art approaches have introduced visual information for speech enhancement,since the visual aspect of speech is essentially unaffected…

Audio and Speech Processing · Electrical Eng. & Systems 2022-04-21 Xinmeng Xu , Yang Wang , Dongxiang Xu , Yiyuan Peng , Cong Zhang , Jie Jia , Binbin Chen

Diffusion models have recently achieved remarkable success in generative modeling, yet their training dynamics across different noise levels remain highly imbalanced, which can lead to inefficient optimization and unstable learning…

Machine Learning · Computer Science 2026-03-12 Nanlong Sun , Lei Shi

Generative speech enhancement methods based on generative adversarial networks (GANs) and diffusion models have shown promising results in various speech enhancement tasks. However, their performance in very low signal-to-noise ratio (SNR)…

Audio and Speech Processing · Electrical Eng. & Systems 2025-08-29 Shrishti Saha Shetu , Emanuël A. P. Habets , Andreas Brendel

In this paper, the coherence between speech and noise signals is used to obtain a Speech Enhancement (SE) gain function, in combination with a Super Gaussian Joint Maximum a Posteriori (SGJMAP) single microphone SE gain function. The…

Audio and Speech Processing · Electrical Eng. & Systems 2020-01-28 Nikhil Shankar , Gautam S Bhat , Chandan K A Reddy , Issa Panahi

In this paper, a speech enhancement method based on noise compensation performed on short time magnitude as well phase spectra is presented. Unlike the conventional geometric approach (GA) to spectral subtraction (SS), here the noise…

Audio and Speech Processing · Electrical Eng. & Systems 2018-03-09 Md Tauhidul Islam , Udoy Saha , K. T. Shahid , Ahmed Bin Hussain , Celia Shahnaz

Multichannel speech enhancement algorithms are essential for improving the intelligibility of speech signals in noisy environments. These algorithms are usually evaluated at the utterance level, but this approach overlooks the disparities…

Sound · Computer Science 2025-06-24 Nasser-Eddine Monir , Paul Magron , Romain Serizel

In this work, we investigate application of generative speech enhancement to improve the robustness of ASR models in noisy and reverberant conditions. We employ a recently-proposed speech enhancement model based on Schr\"odinger bridge,…

Audio and Speech Processing · Electrical Eng. & Systems 2025-05-09 Rauf Nasretdinov , Roman Korostik , Ante Jukić

We address the problem of speech enhancement generalisation to unseen environments by performing two manipulations. First, we embed an additional recording from the environment alone, and use this embedding to alter activations in the main…

Audio and Speech Processing · Electrical Eng. & Systems 2018-10-31 Gil Keren , Jing Han , Björn Schuller

Diffusion-based speech enhancement (SE) achieves natural-sounding speech and strong generalization, yet suffers from key limitations like generative artifacts and high inference latency. In this work, we systematically study artifact…

We enhance the vanilla adversarial training method for unsupervised Automatic Speech Recognition (ASR) by a diffusion-GAN. Our model (1) injects instance noises of various intensities to the generator's output and unlabeled reference text…

Computation and Language · Computer Science 2023-03-27 Xianchao Wu

Score-based diffusion models are a class of generative models whose dynamics is described by stochastic differential equations that map noise into data. While recent works have started to lay down a theoretical foundation for these models,…

A two-step enhancement method based on spectral subtraction and phase spectrum compensation is presented in this paper for noisy speeches in adverse environments involving non-stationary noise and medium to low levels of SNR. The magnitude…

Audio and Speech Processing · Electrical Eng. & Systems 2018-03-02 Md Tauhidul Islam , Asaduzzaman , Celia Shahnaz , Wei-Ping Zhu , M. Omair Ahmad

Stochastic averaging allows for the reduction of the dimension and complexity of stochastic dynamical systems with multiple time scales, replacing fast variables with statistically equivalent stochastic processes in order to analyze…

Probability · Mathematics 2015-02-25 William F. Thompson , Rachel A. Kuske , Adam H. Monahan
‹ Prev 1 3 4 5 6 7 10 Next ›