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

Text-to-Speech synthesis systems are generally evaluated using Mean OpinionScore (MOS) tests, where listeners score samples of synthetic speech on aLikert scale. A major drawback of MOS tests is that they only offer a generalmeasure of overall quality-i.e., the naturalness of an utterance-and so cannottell us where exactly synthesis errors occur. This can make evaluation of theappropriateness of prosodic variation within utterances inconclusive. Toaddress this, we propose a novel evaluation method based on the Rapid ProsodyTranscription paradigm. This allows listeners to mark the locations of errorsin an utterance in real-time, providing a probabilistic representation of theperceptual errors that occur in the synthetic signal. We conduct experimentsthat confirm that the fine-grained evaluation can be mapped to system rankingsof standard MOS tests, but the error marking gives a much more comprehensiveassessment of synthesized prosody. In particular, for standard audiobook testset samples, we see that error marks consistently cluster around words at majorprosodic boundaries indicated by punctuation. However, for question-answerbased stimuli, where we control information structure, we see differencesemerge in the ability of neural TTS systems to generate context-appropriateprosodic prominence.

Location, Location: Enhancing the Evaluation of Text-to-Speech Synthesis Using the Rapid Prosody Transcription Paradigm