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

Target speech separation is the process of filtering a certain speaker'svoice out of speech mixtures according to the additional speaker identityinformation provided. Recent works have made considerable improvement byprocessing signals in the time domain directly. The majority of them take fullyoverlapped speech mixtures for training. However, since most real-lifeconversations occur randomly and are sparsely overlapped, we argue thattraining with different overlap ratio data benefits. To do so, an unavoidableproblem is that the popularly used SI-SNR loss has no definition for silentsources. This paper proposes the weighted SI-SNR loss, together with the jointlearning of target speech separation and personal VAD. The weighted SI-SNR lossimposes a weight factor that is proportional to the target speaker's durationand returns zero when the target speaker is absent. Meanwhile, the personal VADgenerates masks and sets non-target speech to silence. Experiments show thatour proposed method outperforms the baseline by 1.73 dB in terms of SDR onfully overlapped speech, as well as by 4.17 dB and 0.9 dB on sparselyoverlapped speech of clean and noisy conditions. Besides, with slightdegradation in performance, our model could reduce the time costs in inference.

Sparsely Overlapped Speech Training in the Time Domain: Joint Learning of Target Speech Separation and Personal VAD Benefits