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

In a multi-channel separation task with multiple speakers, we aim to recoverall individual speech signals from the mixture. In contrast to single-channelapproaches, which rely on the different spectro-temporal characteristics of thespeech signals, multi-channel approaches should additionally utilize thedifferent spatial locations of the sources for a more powerful separationespecially when the number of sources increases. To enhance the spatialprocessing in a multi-channel source separation scenario, in this work, wepropose a deep neural network (DNN) based spatially selective filter (SSF) thatcan be spatially steered to extract the speaker of interest by initializing arecurrent neural network layer with the target direction. We compare theproposed SSF with a common end-to-end direct separation (DS) approach trainedusing utterance-wise permutation invariant training (PIT), which onlyimplicitly learns to perform spatial filtering. We show that the SSF has aclear advantage over a DS approach with the same underlying networkarchitecture when there are more than two speakers in the mixture, which can beattributed to a better use of the spatial information. Furthermore, we findthat the SSF generalizes much better to additional noise sources that were notseen during training and to scenarios with speakers positioned at a similarangle.

Multi-channel Speech Separation Using Spatially Selective Deep Non-linear Filters