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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Improving text representation has attracted much attention to achieveexpressive text-to-speech (TTS). However, existing works only implicitly learnthe prosody with masked token reconstruction tasks, which leads to low trainingefficiency and difficulty in prosody modeling. We propose CLAPSpeech, across-modal contrastive pre-training framework that explicitly learns theprosody variance of the same text token under different contexts. Specifically,1) We encourage the model to connect the text context with its correspondingprosody pattern in the joint multi-modal space with the elaborate design of theencoder inputs and contrastive loss; 2) We introduce a multi-scale pre-trainingpipeline to capture prosody patterns in multiple levels. We show how toincorporate CLAPSpeech into existing TTS models for better prosody. Experimentson three datasets not only show that CLAPSpeech could improve the prosodyprediction for existing TTS methods, but also demonstrate its generalizationability to adapt to multiple languages and multi-speaker TTS. We also deeplyanalyze the principle behind the performance of CLAPSpeech. Ablation studiesdemonstrate the necessity of each component in our method. Source code andaudio samples are available at https://clapspeech.github.io.

CLAPSpeech: Learning Prosody from Text Context with Contrastive Language-Audio Pre-training