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/

Zendy Plus

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

Zendy Tools

Zendy Open

In this paper, we propose two mask-based beamforming methods using a deepneural network (DNN) trained by multichannel loss functions. Beamformingtechnique using time-frequency (TF)-masks estimated by a DNN have been appliedto many applications where TF-masks are used for estimating spatial covariancematrices. To train a DNN for mask-based beamforming, loss functions designedfor monaural speech enhancement/separation have been employed. Although such atraining criterion is simple, it does not directly correspond to theperformance of mask-based beamforming. To overcome this problem, we usemultichannel loss functions which evaluate the estimated spatial covariancematrices based on the multichannel Itakura--Saito divergence. DNNs trained bythe multichannel loss functions can be applied to construct severalbeamformers. Experimental results confirmed their effectiveness and robustnessto microphone configurations.

Multichannel Loss Function for Supervised Speech Source Separation by Mask-based Beamforming