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

Nowadays, recognition-synthesis-based methods have been quite popular withvoice conversion (VC). By introducing linguistics features with gooddisentangling characters extracted from an automatic speech recognition (ASR)model, the VC performance achieved considerable breakthroughs. Recently,self-supervised learning (SSL) methods trained with a large-scale unannotatedspeech corpus have been applied to downstream tasks focusing on the contentinformation, which is suitable for VC tasks. However, a huge amount of speakerinformation in SSL representations degrades timbre similarity and the qualityof converted speech significantly. To address this problem, we proposed ahigh-similarity any-to-one voice conversion method with the input of SSLrepresentations. We incorporated adversarial training mechanisms in thesynthesis module using external unannotated corpora. Two auxiliarydiscriminators were trained to distinguish whether a sequence ofmel-spectrograms has been converted by the acoustic model and whether asequence of content embeddings contains speaker information from externalcorpora. Experimental results show that our proposed method achieves comparablesimilarity and higher naturalness than the supervised method, which needs ahuge amount of annotated corpora for training and is applicable to improvesimilarity for VC methods with other SSL representations as input.

Adversarial Speaker Disentanglement Using Unannotated External Data for Self-supervised Representation Based Voice Conversion