English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

This paper describes multichannel speech enhancement for improving automaticspeech recognition (ASR) in noisy environments. Recently, the minimum variancedistortionless response (MVDR) beamforming has widely been used because itworks well if the steering vector of speech and the spatial covariance matrix(SCM) of noise are given. To estimating such spatial information, conventionalstudies take a supervised approach that classifies each time-frequency (TF) bininto noise or speech by training a deep neural network (DNN). The performanceof ASR, however, is degraded in an unknown noisy environment. To solve thisproblem, we take an unsupervised approach that decomposes each TF bin into thesum of speech and noise by using multichannel nonnegative matrix factorization(MNMF). This enables us to accurately estimate the SCMs of speech and noise notfrom observed noisy mixtures but from separated speech and noise components. Inthis paper we propose online MVDR beamforming by effectively initializing andincrementally updating the parameters of MNMF. Another main contribution is tocomprehensively investigate the performances of ASR obtained by various typesof spatial filters, i.e., time-invariant and variant versions of MVDRbeamformers and those of rank-1 and full-rank multichannel Wiener filters, incombination with MNMF. The experimental results showed that the proposed methodoutperformed the state-of-the-art DNN-based beamforming method in unknownenvironments that did not match training data.

Unsupervised Speech Enhancement Based on Multichannel NMF-Informed Beamforming for Noise-Robust Automatic Speech Recognition