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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We present a novel multi-channel front-end based on channel shortening withtheWeighted Prediction Error (WPE) method followed by a fixed MVDR beamformerused in combination with a recently proposed self-attention-based channelcombination (SACC) scheme, for tackling the distant ASR problem. We show thatthe proposed system used as part of a ContextNet based end-to-end (E2E) ASRsystem outperforms leading ASR systems as demonstrated by a 21.6% reduction inrelative WER on a multi-channel LibriSpeech playback dataset. We also show howdereverberation prior to beamforming is beneficial and compare the WPE methodwith a modified neural channel shortening approach. An analysis of thenon-intrusive estimate of the signal C50 confirms that the 8 channel WPE methodprovides significant dereverberation of the signals (13.6 dB improvement). Wealso show how the weights of the SACC system allow the extraction of accuratespatial information which can be beneficial for other speech processingapplications like diarization.

Spatial Processing Front-End For Distant ASR Exploiting Self-Attention Channel Combinator