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

With the development of deep learning, neural network-based speechenhancement (SE) models have shown excellent performance. Meanwhile, it wasshown that the development of self-supervised pre-trained models can be appliedto various downstream tasks. In this paper, we will consider the application ofthe pre-trained model to the real-time SE problem. Specifically, the encoderand bottleneck layer of the DEMUCS model are initialized using theself-supervised pretrained WavLM model, the convolution in the encoder isreplaced by causal convolution, and the transformer encoder in the bottlenecklayer is based on causal attention mask. In addition, as discretizing the noisyspeech representations is more beneficial for denoising, we utilize aquantization module to discretize the representation output from the bottlenecklayer, which is then fed into the decoder to reconstruct the clean speechwaveform. Experimental results on the Valentini dataset and an internal datasetshow that the pre-trained model based initialization can improve the SEperformance and the discretization operation suppresses the noise component inthe representations to some extent, which can further improve the performance.

Speech Enhancement Using Self-Supervised Pre-Trained Model and Vector Quantization