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

We are interested in a challenging task, Realistic-Music-Score based SingingVoice Synthesis (RMS-SVS). RMS-SVS aims to generate high-quality singing voicesgiven realistic music scores with different note types (grace, slur, rest,etc.). Though significant progress has been achieved, recent singing voicesynthesis (SVS) methods are limited to fine-grained music scores, which requirea complicated data collection pipeline with time-consuming manual annotation toalign music notes with phonemes. Furthermore, these manual annotation destroysthe regularity of note durations in music scores, making fine-grained musicscores inconvenient for composing. To tackle these challenges, we proposeRMSSinger, the first RMS-SVS method, which takes realistic music scores asinput, eliminating most of the tedious manual annotation and avoiding theaforementioned inconvenience. Note that music scores are based on words ratherthan phonemes, in RMSSinger, we introduce word-level modeling to avoid thetime-consuming phoneme duration annotation and the complicated phoneme-levelmel-note alignment. Furthermore, we propose the first diffusion-based pitchmodeling method, which ameliorates the naturalness of existing pitch-modelingmethods. To achieve these, we collect a new dataset containing realistic musicscores and singing voices according to these realistic music scores fromprofessional singers. Extensive experiments on the dataset demonstrate theeffectiveness of our methods. Audio samples are available athttps://rmssinger.github.io/.

RMSSinger: Realistic-Music-Score based Singing Voice Synthesis