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

Language models (LMs) have shown superior performances in various speechgeneration tasks recently, demonstrating their powerful ability for semanticcontext modeling. Given the intrinsic similarity between speech generation andspeech enhancement, harnessing semantic information holds potential advantagesfor speech enhancement tasks. In light of this, we propose SELM, a novelparadigm for speech enhancement, which integrates discrete tokens and leverageslanguage models. SELM comprises three stages: encoding, modeling, and decoding.We transform continuous waveform signals into discrete tokens using pre-trainedself-supervised learning (SSL) models and a k-means tokenizer. Language modelsthen capture comprehensive contextual information within these tokens. Finally,a detokenizer and HiFi-GAN restore them into enhanced speech. Experimentalresults demonstrate that SELM achieves comparable performance in objectivemetrics alongside superior results in subjective perception. Our demos areavailable https://honee-w.github.io/SELM/.

SELM: Speech Enhancement Using Discrete Tokens and Language Models